test_anonymization.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Anonymization examples\n",
    "### References:\n",
    "#### [1] Josep Domingo-Ferrer and Vicenç Torra, \"Ordinal, continuous and heterogeneous k-anonymity through microaggregation\", Data Mining and Knowledge Discovery, Vol. 11, pp. 195-212, Sep 2005. DOI: https://doi.org/10.1007/s10618-005-0007-5\n",
    "#### [2] Jordi Soria-Comas, Josep Domingo-Ferrer, David Sánchez and Sergio Martínez, \"t-Closeness through microaggregation: strict privacy with enhanced utility preservation\", IEEE Transactions on Knowledge and Data Engineering, Vol. 27, no. 11, pp. 3098-3110, Oct 2015. DOI: https://doi.org/10.1109/TKDE.2015.2435777\n",
    "#### [3] Jordi Soria-Comas, Josep Domingo-Ferrer, David Sánchez and Sergio Martínez, \"Enhancing data utility in differential privacy via microaggregation-based k-anonymity\", The VLDB Journal, Vol. 23, no. 5, pp. 771-794, Sep 2014. DOI: https://doi.org/10.1007/s00778-014-0351-4\n",
    "#### [4] Josep Domingo-Ferrer and Vicenç Torra, \"Disclosure risk assessment in statistical data protection\", Journal of Computational and Applied Mathematics, Vol. 164, pp. 285-293, Mar 2004. DOI: https://doi.org/10.1016/S0377-0427(03)00643-5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. k-Anomymity via MDAV"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Toy data set that includes all available attribute types"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# pip install -i https://test.pypi.org/pypi/ --extra-index-url https://pypi.org/simple anonymization-crisesURV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# !pip install -i https://test.pypi.org/simple/ anonymization-crisesURV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "from anonymization.src.entities.dataset_CSV import Dataset_CSV\n",
    "from anonymization.src.entities.dataset_DataFrame import Dataset_DataFrame\n",
    "from anonymization.src.algorithms.anonymization_scheme import Anonymization_scheme\n",
    "from anonymization.src.algorithms.k_anonymity import K_anonymity\n",
    "from anonymization.src.algorithms.mdav import Mdav\n",
    "from anonymization.src.utils import utils"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Following, it is indicated the path to the csv file containing the data set and the path to the xml file describing the attributes in the data set. Inside the xml file, there is a detailed descritpion about how to fill this xml file in order to properly configure the different attribute types in the data set  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# path_csv = \"https://raw.github.com/CrisesUrv/SoBigDataTraining/master/anonymization/input_datasets/toy_all_types.csv\"\n",
    "# path_settings = \"https://raw.github.com/CrisesUrv/SoBigDataTraining/master/anonymization/input_datasets/metadata_toy_all_types.xml\"\n",
    "path_csv = \"anonymization/input_datasets/toy_all_types.csv\"\n",
    "path_settings = \"anonymization/input_datasets/metadata_toy_all_types.xml\"\n",
    "data_frame = utils.read_dataframe_from_csv(path_csv)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1.1 The data set is loaded from a DataFrame passed as parameter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading dataset\n",
      "Dataset loaded: anonymization/input_datasets/toy_all_types.csv\n",
      "Records loaded: 10\n",
      "Dataset: anonymization/input_datasets/toy_all_types.csv\n",
      "Dataset head:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>date</th>\n",
       "      <th>occupation</th>\n",
       "      <th>native-country</th>\n",
       "      <th>location</th>\n",
       "      <th>datetime</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>40</td>\n",
       "      <td>39.5</td>\n",
       "      <td>23546</td>\n",
       "      <td>1/11/2016</td>\n",
       "      <td>clerk</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.8430139:10.507994</td>\n",
       "      <td>2011-02-03 08:34:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>13</td>\n",
       "      <td>50.3</td>\n",
       "      <td>10230</td>\n",
       "      <td>6/12/2015</td>\n",
       "      <td>executive</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.54427:10.32615</td>\n",
       "      <td>2011-02-03 09:34:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>40</td>\n",
       "      <td>38.0</td>\n",
       "      <td>0</td>\n",
       "      <td>19/7/2015</td>\n",
       "      <td>cleaner</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.70853:10.4036</td>\n",
       "      <td>2011-02-03 10:34:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>40</td>\n",
       "      <td>53.1</td>\n",
       "      <td>152000</td>\n",
       "      <td>25/7/2015</td>\n",
       "      <td>cleaner</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.77925:11.24626</td>\n",
       "      <td>2011-02-04 10:34:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>40</td>\n",
       "      <td>28.8</td>\n",
       "      <td>54120</td>\n",
       "      <td>10/8/2016</td>\n",
       "      <td>specialist</td>\n",
       "      <td>Cuba</td>\n",
       "      <td>43.8430139:10.507994</td>\n",
       "      <td>2011-02-04 08:34:04</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   hours-per-week   age  income       date  occupation native-country  \\\n",
       "0              40  39.5   23546  1/11/2016       clerk  United_States   \n",
       "1              13  50.3   10230  6/12/2015   executive  United_States   \n",
       "2              40  38.0       0  19/7/2015     cleaner  United_States   \n",
       "3              40  53.1  152000  25/7/2015     cleaner  United_States   \n",
       "4              40  28.8   54120  10/8/2016  specialist           Cuba   \n",
       "\n",
       "               location             datetime  \n",
       "0  43.8430139:10.507994  2011-02-03 08:34:04  \n",
       "1     43.54427:10.32615  2011-02-03 09:34:04  \n",
       "2      43.70853:10.4036  2011-02-03 10:34:04  \n",
       "3     43.77925:11.24626  2011-02-04 10:34:04  \n",
       "4  43.8430139:10.507994  2011-02-04 08:34:04  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Dataset description:\n",
      "Data set: anonymization/input_datasets/toy_all_types.csv\n",
      "Records: 10\n",
      "Attributes:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Attribute_type</th>\n",
       "      <th>Sensitivity_type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>hours-per-week</td>\n",
       "      <td>numerical_discrete</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>age</td>\n",
       "      <td>numerical_continuous</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>income</td>\n",
       "      <td>numerical_discrete</td>\n",
       "      <td>confidential</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>date</td>\n",
       "      <td>date</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>occupation</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>native-country</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>location</td>\n",
       "      <td>coordinate</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>datetime</td>\n",
       "      <td>datetime</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Name        Attribute_type  Sensitivity_type\n",
       "0  hours-per-week    numerical_discrete  quasi_identifier\n",
       "1             age  numerical_continuous  quasi_identifier\n",
       "2          income    numerical_discrete      confidential\n",
       "3            date                  date  quasi_identifier\n",
       "4      occupation     plain_categorical  quasi_identifier\n",
       "5  native-country     plain_categorical  quasi_identifier\n",
       "6        location            coordinate  quasi_identifier\n",
       "7        datetime              datetime  quasi_identifier"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "dataset = Dataset_DataFrame(data_frame, path_settings)\n",
    "dataset.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1.2 The data set is anonymized, in this case, applying k_anonymity via MDAV with a privacy level of k=3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Anonymizing k-Anonymity, k = 3 via MDAV\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0c97da8bdb7c433587c2c58564f4d988",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=10.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:00\n"
     ]
    }
   ],
   "source": [
    "k = 3\n",
    "anonymization_scheme = K_anonymity(dataset, k)\n",
    "algorithm = Mdav()\n",
    "anonymization_scheme.calculate_anonymization(algorithm)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1.3 Information loss metrics are calculated by comparing original and anonymized data set  \n",
    "Metrics calculated: Sum of Square Error (SSE) and, for each attribute, mean and variance\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Calculating information loss metrics\n",
      "\n",
      "Information loss metrics:\n",
      "SSE: 0.892\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Original mean</th>\n",
       "      <th>Anonymized mean</th>\n",
       "      <th>Original variance</th>\n",
       "      <th>Anonymized variance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>hours-per-week</td>\n",
       "      <td>36.4</td>\n",
       "      <td>36.3</td>\n",
       "      <td>1.302000e+02</td>\n",
       "      <td>5.550000e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>age</td>\n",
       "      <td>42.33</td>\n",
       "      <td>42.34</td>\n",
       "      <td>6.654810e+01</td>\n",
       "      <td>2.932440e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>income</td>\n",
       "      <td>35771.4</td>\n",
       "      <td>35771.4</td>\n",
       "      <td>1.687522e+09</td>\n",
       "      <td>1.687522e+09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>date</td>\n",
       "      <td>18/2/2016</td>\n",
       "      <td>18/2/2016</td>\n",
       "      <td>3.245814e+14</td>\n",
       "      <td>7.524886e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>occupation</td>\n",
       "      <td>executive</td>\n",
       "      <td>specialist</td>\n",
       "      <td>6.000000e-01</td>\n",
       "      <td>6.000000e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>native-country</td>\n",
       "      <td>United_States</td>\n",
       "      <td>United_States</td>\n",
       "      <td>2.000000e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>location</td>\n",
       "      <td>43.70666917:10.495949199999998</td>\n",
       "      <td>43.70666917:10.495949200000002</td>\n",
       "      <td>8.156504e-02</td>\n",
       "      <td>1.272151e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>datetime</td>\n",
       "      <td>2011-02-04 09:46:04</td>\n",
       "      <td>2011-02-04 09:46:04</td>\n",
       "      <td>4.491418e+09</td>\n",
       "      <td>1.714414e+09</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Name                   Original mean  \\\n",
       "0  hours-per-week                            36.4   \n",
       "1             age                           42.33   \n",
       "2          income                         35771.4   \n",
       "3            date                       18/2/2016   \n",
       "4      occupation                       executive   \n",
       "5  native-country                   United_States   \n",
       "6        location  43.70666917:10.495949199999998   \n",
       "7        datetime             2011-02-04 09:46:04   \n",
       "\n",
       "                  Anonymized mean  Original variance  Anonymized variance  \n",
       "0                            36.3       1.302000e+02         5.550000e+01  \n",
       "1                           42.34       6.654810e+01         2.932440e+01  \n",
       "2                         35771.4       1.687522e+09         1.687522e+09  \n",
       "3                       18/2/2016       3.245814e+14         7.524886e+13  \n",
       "4                      specialist       6.000000e-01         6.000000e-01  \n",
       "5                   United_States       2.000000e-01         0.000000e+00  \n",
       "6  43.70666917:10.495949200000002       8.156504e-02         1.272151e-02  \n",
       "7             2011-02-04 09:46:04       4.491418e+09         1.714414e+09  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "information_loss = Anonymization_scheme.calculate_information_loss(dataset, anonymization_scheme.anonymized_dataset)\n",
    "information_loss.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1.4 Disclosure risk is calculated via record linkage between anonymized and original data set  \n",
    "The disclosure risk estimates the percentage of anonymized records that correct match with the original ones"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Calculating record linkage (disclosure risk)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0106e018f7f44a41bf57120d23058d91",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=10.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Disclosure risk: 3.000 (30.00%)\n"
     ]
    }
   ],
   "source": [
    "disclosure_risk = Anonymization_scheme.calculate_record_linkage(dataset, anonymization_scheme.anonymized_dataset)\n",
    "disclosure_risk.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1.5 The anonymized data set can be saved to a csv formated file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Dataset saved: ./anonymization/output_datasets/toy_all_types_anom.csv'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "anonymization_scheme.save_anonymized_dataset(\"./anonymization/output_datasets/toy_all_types_anom.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1.6 The anonymized data set can be converted to DataFrame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>date</th>\n",
       "      <th>occupation</th>\n",
       "      <th>native-country</th>\n",
       "      <th>location</th>\n",
       "      <th>datetime</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>40</td>\n",
       "      <td>38.4</td>\n",
       "      <td>23546</td>\n",
       "      <td>28/4/2016</td>\n",
       "      <td>clerk</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.75335796666667:10.438398000000001</td>\n",
       "      <td>2011-02-03 17:34:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>25</td>\n",
       "      <td>50.6</td>\n",
       "      <td>10230</td>\n",
       "      <td>18/9/2015</td>\n",
       "      <td>executive</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.643851299999994:10.386764666666666</td>\n",
       "      <td>2011-02-04 10:34:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>40</td>\n",
       "      <td>38.4</td>\n",
       "      <td>0</td>\n",
       "      <td>28/4/2016</td>\n",
       "      <td>clerk</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.75335796666667:10.438398000000001</td>\n",
       "      <td>2011-02-03 17:34:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>42</td>\n",
       "      <td>39.1</td>\n",
       "      <td>152000</td>\n",
       "      <td>21/4/2016</td>\n",
       "      <td>specialist</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.718765975:10.621001</td>\n",
       "      <td>2011-02-04 21:19:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>42</td>\n",
       "      <td>39.1</td>\n",
       "      <td>54120</td>\n",
       "      <td>21/4/2016</td>\n",
       "      <td>specialist</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.718765975:10.621001</td>\n",
       "      <td>2011-02-04 21:19:04</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  hours-per-week   age  income       date  occupation native-country  \\\n",
       "0             40  38.4   23546  28/4/2016       clerk  United_States   \n",
       "1             25  50.6   10230  18/9/2015   executive  United_States   \n",
       "2             40  38.4       0  28/4/2016       clerk  United_States   \n",
       "3             42  39.1  152000  21/4/2016  specialist  United_States   \n",
       "4             42  39.1   54120  21/4/2016  specialist  United_States   \n",
       "\n",
       "                                location             datetime  \n",
       "0   43.75335796666667:10.438398000000001  2011-02-03 17:34:04  \n",
       "1  43.643851299999994:10.386764666666666  2011-02-04 10:34:04  \n",
       "2   43.75335796666667:10.438398000000001  2011-02-03 17:34:04  \n",
       "3                 43.718765975:10.621001  2011-02-04 21:19:04  \n",
       "4                 43.718765975:10.621001  2011-02-04 21:19:04  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_anonymized = anonymization_scheme.anonymized_dataset_to_dataframe()\n",
    "df_anonymized.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 1.7 The previously saved anonymized and original data sets can be loaded in order to calculate the privacy metrics a posteriori"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading dataset\n",
      "Dataset loaded: ./anonymization/input_datasets/toy_all_types.csv\n",
      "Records loaded: 10\n",
      "Dataset: ./anonymization/input_datasets/toy_all_types.csv\n",
      "Dataset head:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>date</th>\n",
       "      <th>occupation</th>\n",
       "      <th>native-country</th>\n",
       "      <th>location</th>\n",
       "      <th>datetime</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>40</td>\n",
       "      <td>39.5</td>\n",
       "      <td>23546</td>\n",
       "      <td>1/11/2016</td>\n",
       "      <td>clerk</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.8430139:10.507994</td>\n",
       "      <td>2011-02-03 08:34:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>13</td>\n",
       "      <td>50.3</td>\n",
       "      <td>10230</td>\n",
       "      <td>6/12/2015</td>\n",
       "      <td>executive</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.54427:10.32615</td>\n",
       "      <td>2011-02-03 09:34:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>40</td>\n",
       "      <td>38.0</td>\n",
       "      <td>0</td>\n",
       "      <td>19/7/2015</td>\n",
       "      <td>cleaner</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.70853:10.4036</td>\n",
       "      <td>2011-02-03 10:34:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>40</td>\n",
       "      <td>53.1</td>\n",
       "      <td>152000</td>\n",
       "      <td>25/7/2015</td>\n",
       "      <td>cleaner</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.77925:11.24626</td>\n",
       "      <td>2011-02-04 10:34:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>40</td>\n",
       "      <td>28.8</td>\n",
       "      <td>54120</td>\n",
       "      <td>10/8/2016</td>\n",
       "      <td>specialist</td>\n",
       "      <td>Cuba</td>\n",
       "      <td>43.8430139:10.507994</td>\n",
       "      <td>2011-02-04 08:34:04</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   hours-per-week   age  income       date  occupation native-country  \\\n",
       "0              40  39.5   23546  1/11/2016       clerk  United_States   \n",
       "1              13  50.3   10230  6/12/2015   executive  United_States   \n",
       "2              40  38.0       0  19/7/2015     cleaner  United_States   \n",
       "3              40  53.1  152000  25/7/2015     cleaner  United_States   \n",
       "4              40  28.8   54120  10/8/2016  specialist           Cuba   \n",
       "\n",
       "               location             datetime  \n",
       "0  43.8430139:10.507994  2011-02-03 08:34:04  \n",
       "1     43.54427:10.32615  2011-02-03 09:34:04  \n",
       "2      43.70853:10.4036  2011-02-03 10:34:04  \n",
       "3     43.77925:11.24626  2011-02-04 10:34:04  \n",
       "4  43.8430139:10.507994  2011-02-04 08:34:04  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Dataset description:\n",
      "Data set: ./anonymization/input_datasets/toy_all_types.csv\n",
      "Records: 10\n",
      "Attributes:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Attribute_type</th>\n",
       "      <th>Sensitivity_type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>hours-per-week</td>\n",
       "      <td>numerical_discrete</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>age</td>\n",
       "      <td>numerical_continuous</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>income</td>\n",
       "      <td>numerical_discrete</td>\n",
       "      <td>confidential</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>date</td>\n",
       "      <td>date</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>occupation</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>native-country</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>location</td>\n",
       "      <td>coordinate</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>datetime</td>\n",
       "      <td>datetime</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Name        Attribute_type  Sensitivity_type\n",
       "0  hours-per-week    numerical_discrete  quasi_identifier\n",
       "1             age  numerical_continuous  quasi_identifier\n",
       "2          income    numerical_discrete      confidential\n",
       "3            date                  date  quasi_identifier\n",
       "4      occupation     plain_categorical  quasi_identifier\n",
       "5  native-country     plain_categorical  quasi_identifier\n",
       "6        location            coordinate  quasi_identifier\n",
       "7        datetime              datetime  quasi_identifier"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "path_csv = \"./anonymization/input_datasets/toy_all_types.csv\"\n",
    "path_settings = \"./anonymization/input_datasets/metadata_toy_all_types.xml\"\n",
    "df = utils.read_dataframe_from_csv(path_csv)\n",
    "dataset_original = Dataset_DataFrame(df, path_settings)\n",
    "dataset_original.description()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading dataset\n",
      "Dataset loaded: ./anonymization/output_datasets/toy_all_types_anom.csv\n",
      "Records loaded: 10\n",
      "Dataset: ./anonymization/output_datasets/toy_all_types_anom.csv\n",
      "Dataset head:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>date</th>\n",
       "      <th>occupation</th>\n",
       "      <th>native-country</th>\n",
       "      <th>location</th>\n",
       "      <th>datetime</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>40</td>\n",
       "      <td>38.4</td>\n",
       "      <td>23546</td>\n",
       "      <td>28/4/2016</td>\n",
       "      <td>clerk</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.75335796666667:10.438398000000001</td>\n",
       "      <td>2011-02-03 17:34:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>25</td>\n",
       "      <td>50.6</td>\n",
       "      <td>10230</td>\n",
       "      <td>18/9/2015</td>\n",
       "      <td>executive</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.643851299999994:10.386764666666666</td>\n",
       "      <td>2011-02-04 10:34:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>40</td>\n",
       "      <td>38.4</td>\n",
       "      <td>0</td>\n",
       "      <td>28/4/2016</td>\n",
       "      <td>clerk</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.75335796666667:10.438398000000001</td>\n",
       "      <td>2011-02-03 17:34:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>42</td>\n",
       "      <td>39.1</td>\n",
       "      <td>152000</td>\n",
       "      <td>21/4/2016</td>\n",
       "      <td>specialist</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.718765975:10.621001</td>\n",
       "      <td>2011-02-04 21:19:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>42</td>\n",
       "      <td>39.1</td>\n",
       "      <td>54120</td>\n",
       "      <td>21/4/2016</td>\n",
       "      <td>specialist</td>\n",
       "      <td>United_States</td>\n",
       "      <td>43.718765975:10.621001</td>\n",
       "      <td>2011-02-04 21:19:04</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   hours-per-week   age  income       date  occupation native-country  \\\n",
       "0              40  38.4   23546  28/4/2016       clerk  United_States   \n",
       "1              25  50.6   10230  18/9/2015   executive  United_States   \n",
       "2              40  38.4       0  28/4/2016       clerk  United_States   \n",
       "3              42  39.1  152000  21/4/2016  specialist  United_States   \n",
       "4              42  39.1   54120  21/4/2016  specialist  United_States   \n",
       "\n",
       "                                location             datetime  \n",
       "0   43.75335796666667:10.438398000000001  2011-02-03 17:34:04  \n",
       "1  43.643851299999994:10.386764666666666  2011-02-04 10:34:04  \n",
       "2   43.75335796666667:10.438398000000001  2011-02-03 17:34:04  \n",
       "3                 43.718765975:10.621001  2011-02-04 21:19:04  \n",
       "4                 43.718765975:10.621001  2011-02-04 21:19:04  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Dataset description:\n",
      "Data set: ./anonymization/output_datasets/toy_all_types_anom.csv\n",
      "Records: 10\n",
      "Attributes:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Attribute_type</th>\n",
       "      <th>Sensitivity_type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>hours-per-week</td>\n",
       "      <td>numerical_discrete</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>age</td>\n",
       "      <td>numerical_continuous</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>income</td>\n",
       "      <td>numerical_discrete</td>\n",
       "      <td>confidential</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>date</td>\n",
       "      <td>date</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>occupation</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>native-country</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>location</td>\n",
       "      <td>coordinate</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>datetime</td>\n",
       "      <td>datetime</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Name        Attribute_type  Sensitivity_type\n",
       "0  hours-per-week    numerical_discrete  quasi_identifier\n",
       "1             age  numerical_continuous  quasi_identifier\n",
       "2          income    numerical_discrete      confidential\n",
       "3            date                  date  quasi_identifier\n",
       "4      occupation     plain_categorical  quasi_identifier\n",
       "5  native-country     plain_categorical  quasi_identifier\n",
       "6        location            coordinate  quasi_identifier\n",
       "7        datetime              datetime  quasi_identifier"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "path_csv = \"./anonymization/output_datasets/toy_all_types_anom.csv\"\n",
    "path_settings = \"./anonymization/input_datasets/metadata_toy_all_types.xml\"\n",
    "df = utils.read_dataframe_from_csv(path_csv)\n",
    "dataset_anomymized = Dataset_DataFrame(df, path_settings)\n",
    "dataset_anomymized.description()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Calculating information loss metrics\n",
      "\n",
      "Information loss metrics:\n",
      "SSE: 0.892\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Original mean</th>\n",
       "      <th>Anonymized mean</th>\n",
       "      <th>Original variance</th>\n",
       "      <th>Anonymized variance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>hours-per-week</td>\n",
       "      <td>36.4</td>\n",
       "      <td>36.3</td>\n",
       "      <td>1.302000e+02</td>\n",
       "      <td>5.550000e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>age</td>\n",
       "      <td>42.33</td>\n",
       "      <td>42.34</td>\n",
       "      <td>6.654810e+01</td>\n",
       "      <td>2.932440e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>income</td>\n",
       "      <td>35771.4</td>\n",
       "      <td>35771.4</td>\n",
       "      <td>1.687522e+09</td>\n",
       "      <td>1.687522e+09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>date</td>\n",
       "      <td>18/2/2016</td>\n",
       "      <td>18/2/2016</td>\n",
       "      <td>3.245814e+14</td>\n",
       "      <td>7.524886e+13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>occupation</td>\n",
       "      <td>executive</td>\n",
       "      <td>specialist</td>\n",
       "      <td>6.000000e-01</td>\n",
       "      <td>6.000000e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>native-country</td>\n",
       "      <td>United_States</td>\n",
       "      <td>United_States</td>\n",
       "      <td>2.000000e-01</td>\n",
       "      <td>0.000000e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>location</td>\n",
       "      <td>43.70666917:10.495949199999998</td>\n",
       "      <td>43.70666917:10.495949200000002</td>\n",
       "      <td>8.156504e-02</td>\n",
       "      <td>1.272151e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>datetime</td>\n",
       "      <td>2011-02-04 09:46:04</td>\n",
       "      <td>2011-02-04 09:46:04</td>\n",
       "      <td>4.491418e+09</td>\n",
       "      <td>1.714414e+09</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Name                   Original mean  \\\n",
       "0  hours-per-week                            36.4   \n",
       "1             age                           42.33   \n",
       "2          income                         35771.4   \n",
       "3            date                       18/2/2016   \n",
       "4      occupation                       executive   \n",
       "5  native-country                   United_States   \n",
       "6        location  43.70666917:10.495949199999998   \n",
       "7        datetime             2011-02-04 09:46:04   \n",
       "\n",
       "                  Anonymized mean  Original variance  Anonymized variance  \n",
       "0                            36.3       1.302000e+02         5.550000e+01  \n",
       "1                           42.34       6.654810e+01         2.932440e+01  \n",
       "2                         35771.4       1.687522e+09         1.687522e+09  \n",
       "3                       18/2/2016       3.245814e+14         7.524886e+13  \n",
       "4                      specialist       6.000000e-01         6.000000e-01  \n",
       "5                   United_States       2.000000e-01         0.000000e+00  \n",
       "6  43.70666917:10.495949200000002       8.156504e-02         1.272151e-02  \n",
       "7             2011-02-04 09:46:04       4.491418e+09         1.714414e+09  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Calculating record linkage (disclosure risk)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ceb8a89ca74f46c2861d104a404cfba5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=10.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Disclosure risk: 3.000 (30.00%)\n"
     ]
    }
   ],
   "source": [
    "information_loss = Anonymization_scheme.calculate_information_loss(dataset_original, dataset_anomymized)\n",
    "information_loss.description()\n",
    "disclosure_risk = Anonymization_scheme.calculate_record_linkage(dataset_original, dataset_anomymized)\n",
    "disclosure_risk.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "========================================================================================================================\n",
    "## 2. k-Anomymity via MDAV\n",
    "### References:\n",
    "#### [1] Josep Domingo-Ferrer and Vicenç Torra, \"Ordinal, continuous and heterogeneous k-anonymity through microaggregation\", Data Mining and Knowledge Discovery, Vol. 11, pp. 195-212, Sep 2005. DOI: https://doi.org/10.1007/s10618-005-0007-5\n",
    "#### [4] Josep Domingo-Ferrer and Vicenç Torra, \"Disclosure risk assessment in statistical data protection\", Journal of Computational and Applied Mathematics, Vol. 164, pp. 285-293, Mar 2004. DOI: https://doi.org/10.1016/S0377-0427(03)00643-5\n",
    "### Adult data set (45222 records)\n",
    "#### k-anonymity via MDAV (k=10) takes long time (~1h) however, the anonymized data set maintains high utility (SSE = ~0.40)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "from anonymization.src.entities.dataset_DataFrame import Dataset_DataFrame\n",
    "from anonymization.src.algorithms.anonymization_scheme import Anonymization_scheme\n",
    "from anonymization.src.algorithms.k_anonymity import K_anonymity\n",
    "from anonymization.src.algorithms.mdav import Mdav\n",
    "from anonymization.src.utils import utils"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Following, it is indicated the path to the csv file containing the data set and the path to the xml file describing the attributes in the data set. Inside the xml file, there is a detailed descritpion about how to fill this xml file in order to properly configure the different attribute types in the data set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "path_csv = \"./anonymization/input_datasets/adult.csv\"\n",
    "path_settings = \"./anonymization/input_datasets/metadata_adult.xml\"\n",
    "data_frame = utils.read_dataframe_from_csv(path_csv)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.1 The data set is loaded from a DataFrame passed as parameter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading dataset\n",
      "Dataset loaded: ./anonymization/input_datasets/adult.csv\n",
      "Records loaded: 45222\n",
      "Dataset: ./anonymization/input_datasets/adult.csv\n",
      "Dataset head:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>education</th>\n",
       "      <th>marital-status</th>\n",
       "      <th>occupation</th>\n",
       "      <th>relationship</th>\n",
       "      <th>race</th>\n",
       "      <th>sex</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>native-country</th>\n",
       "      <th>prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>25</td>\n",
       "      <td>Private</td>\n",
       "      <td>11th</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Machine-op-inspct</td>\n",
       "      <td>Own-child</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>38</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Farming-fishing</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>50</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>28</td>\n",
       "      <td>Local-gov</td>\n",
       "      <td>Assoc-acdm</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Protective-serv</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>44</td>\n",
       "      <td>Private</td>\n",
       "      <td>Some-college</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Machine-op-inspct</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>34</td>\n",
       "      <td>Private</td>\n",
       "      <td>10th</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Other-service</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>30</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age  workclass     education      marital-status         occupation  \\\n",
       "0   25    Private          11th       Never-married  Machine-op-inspct   \n",
       "1   38    Private       HS-grad  Married-civ-spouse    Farming-fishing   \n",
       "2   28  Local-gov    Assoc-acdm  Married-civ-spouse    Protective-serv   \n",
       "3   44    Private  Some-college  Married-civ-spouse  Machine-op-inspct   \n",
       "4   34    Private          10th       Never-married      Other-service   \n",
       "\n",
       "    relationship   race   sex  hours-per-week native-country prediction  \n",
       "0      Own-child  Black  Male              40  United-States      <=50K  \n",
       "1        Husband  White  Male              50  United-States      <=50K  \n",
       "2        Husband  White  Male              40  United-States       >50K  \n",
       "3        Husband  Black  Male              40  United-States       >50K  \n",
       "4  Not-in-family  White  Male              30  United-States      <=50K  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Dataset description:\n",
      "Data set: ./anonymization/input_datasets/adult.csv\n",
      "Records: 45222\n",
      "Attributes:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Attribute_type</th>\n",
       "      <th>Sensitivity_type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>age</td>\n",
       "      <td>numerical_discrete</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>workclass</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>education</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>marital-status</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>occupation</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>relationship</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>race</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>sex</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>hours-per-week</td>\n",
       "      <td>numerical_discrete</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>native-country</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>prediction</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>confidential</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Name      Attribute_type  Sensitivity_type\n",
       "0              age  numerical_discrete  quasi_identifier\n",
       "1        workclass   plain_categorical  quasi_identifier\n",
       "2        education   plain_categorical  quasi_identifier\n",
       "3   marital-status   plain_categorical  quasi_identifier\n",
       "4       occupation   plain_categorical  quasi_identifier\n",
       "5     relationship   plain_categorical  quasi_identifier\n",
       "6             race   plain_categorical  quasi_identifier\n",
       "7              sex   plain_categorical  quasi_identifier\n",
       "8   hours-per-week  numerical_discrete  quasi_identifier\n",
       "9   native-country   plain_categorical  quasi_identifier\n",
       "10      prediction   plain_categorical      confidential"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "dataset = Dataset_DataFrame(data_frame, path_settings)\n",
    "dataset.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.2 The data set is anonymized, in this case, applying k_anonymity via MDAV with a privacy level of k=10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Anonymizing k-Anonymity, k = 10 via MDAV\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "04c1fa040f134341ad1ff277cfdce91f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 1:00:33\n"
     ]
    }
   ],
   "source": [
    "k = 10\n",
    "anonymization_scheme = K_anonymity(dataset, k)\n",
    "algorithm = Mdav()\n",
    "anonymization_scheme.calculate_anonymization(algorithm)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.3 Information loss metrics are calculated by comparing original and anonymized data set  \n",
    "Metrics calculated: Sum of Square Error (SSE) and, for each attribute, mean and variance\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Calculating information loss metrics\n",
      "\n",
      "Information loss metrics:\n",
      "SSE: 0.418\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Original mean</th>\n",
       "      <th>Anonymized mean</th>\n",
       "      <th>Original variance</th>\n",
       "      <th>Anonymized variance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>age</td>\n",
       "      <td>38.5479</td>\n",
       "      <td>38.5484</td>\n",
       "      <td>174.912587</td>\n",
       "      <td>150.149043</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>workclass</td>\n",
       "      <td>Private</td>\n",
       "      <td>Private</td>\n",
       "      <td>0.263478</td>\n",
       "      <td>0.273318</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>education</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>0.673102</td>\n",
       "      <td>0.681306</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>marital-status</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>0.534408</td>\n",
       "      <td>0.538455</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>occupation</td>\n",
       "      <td>Craft-repair</td>\n",
       "      <td>Prof-specialty</td>\n",
       "      <td>0.866879</td>\n",
       "      <td>0.866437</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>relationship</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Husband</td>\n",
       "      <td>0.587236</td>\n",
       "      <td>0.591084</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>race</td>\n",
       "      <td>White</td>\n",
       "      <td>White</td>\n",
       "      <td>0.139733</td>\n",
       "      <td>0.135332</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>sex</td>\n",
       "      <td>Male</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.324952</td>\n",
       "      <td>0.335235</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>hours-per-week</td>\n",
       "      <td>40.938</td>\n",
       "      <td>40.9396</td>\n",
       "      <td>144.180908</td>\n",
       "      <td>123.105391</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>native-country</td>\n",
       "      <td>United-States</td>\n",
       "      <td>United-States</td>\n",
       "      <td>0.086905</td>\n",
       "      <td>0.039583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>prediction</td>\n",
       "      <td>&lt;=50K</td>\n",
       "      <td>&lt;=50K</td>\n",
       "      <td>0.247844</td>\n",
       "      <td>0.247844</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Name       Original mean     Anonymized mean  Original variance  \\\n",
       "0              age             38.5479             38.5484         174.912587   \n",
       "1        workclass             Private             Private           0.263478   \n",
       "2        education             HS-grad             HS-grad           0.673102   \n",
       "3   marital-status  Married-civ-spouse  Married-civ-spouse           0.534408   \n",
       "4       occupation        Craft-repair      Prof-specialty           0.866879   \n",
       "5     relationship             Husband             Husband           0.587236   \n",
       "6             race               White               White           0.139733   \n",
       "7              sex                Male                Male           0.324952   \n",
       "8   hours-per-week              40.938             40.9396         144.180908   \n",
       "9   native-country       United-States       United-States           0.086905   \n",
       "10      prediction               <=50K               <=50K           0.247844   \n",
       "\n",
       "    Anonymized variance  \n",
       "0            150.149043  \n",
       "1              0.273318  \n",
       "2              0.681306  \n",
       "3              0.538455  \n",
       "4              0.866437  \n",
       "5              0.591084  \n",
       "6              0.135332  \n",
       "7              0.335235  \n",
       "8            123.105391  \n",
       "9              0.039583  \n",
       "10             0.247844  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "information_loss = Anonymization_scheme.calculate_information_loss(dataset, anonymization_scheme.anonymized_dataset)\n",
    "information_loss.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.4 Disclosure risk is calculated via record linkage between anonymized and original data set  \n",
    "The disclosure risk estimates the percentage of anonymized records that correct match with the original ones"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b5997d7cbdba4e168ca01ac2c4175068",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Disclosure risk: 1884.426 (4.17%)\n"
     ]
    }
   ],
   "source": [
    "disclosure_risk = Anonymization_scheme.calculate_fast_record_linkage(dataset, anonymization_scheme.anonymized_dataset)\n",
    "disclosure_risk.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.5 The anonymized data set can be saved to a csv formated file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Dataset saved: ./anonymization/output_datasets/adult_anom.csv'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "anonymization_scheme.save_anonymized_dataset(\"./anonymization/output_datasets/adult_anom.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 2.6 Also, the anonymized data set can be converted to DataFrame (to use as input data for other libraries)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>education</th>\n",
       "      <th>marital-status</th>\n",
       "      <th>occupation</th>\n",
       "      <th>relationship</th>\n",
       "      <th>race</th>\n",
       "      <th>sex</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>native-country</th>\n",
       "      <th>prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>29</td>\n",
       "      <td>Self-emp-not-inc</td>\n",
       "      <td>11th</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Farming-fishing</td>\n",
       "      <td>Own-child</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>28</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Transport-moving</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>45</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>33</td>\n",
       "      <td>Local-gov</td>\n",
       "      <td>Assoc-acdm</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Protective-serv</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>41</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>47</td>\n",
       "      <td>State-gov</td>\n",
       "      <td>Some-college</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Machine-op-inspct</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>42</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>23</td>\n",
       "      <td>Private</td>\n",
       "      <td>10th</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Other-service</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>30</td>\n",
       "      <td>El-Salvador</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  age         workclass     education      marital-status         occupation  \\\n",
       "0  29  Self-emp-not-inc          11th       Never-married    Farming-fishing   \n",
       "1  28           Private       HS-grad  Married-civ-spouse   Transport-moving   \n",
       "2  33         Local-gov    Assoc-acdm  Married-civ-spouse    Protective-serv   \n",
       "3  47         State-gov  Some-college  Married-civ-spouse  Machine-op-inspct   \n",
       "4  23           Private          10th       Never-married      Other-service   \n",
       "\n",
       "    relationship   race   sex hours-per-week native-country prediction  \n",
       "0      Own-child  Black  Male             40  United-States      <=50K  \n",
       "1        Husband  White  Male             45  United-States      <=50K  \n",
       "2        Husband  White  Male             41  United-States       >50K  \n",
       "3        Husband  Black  Male             42  United-States       >50K  \n",
       "4  Not-in-family  White  Male             30    El-Salvador      <=50K  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_anonymized = anonymization_scheme.anonymized_dataset_to_dataframe()\n",
    "df_anonymized.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "========================================================================================================================\n",
    "## 3. k-Anomymity via Microaggregation\n",
    "#### [1] Josep Domingo-Ferrer and Vicenç Torra, \"Ordinal, continuous and heterogeneous k-anonymity through microaggregation\", Data Mining and Knowledge Discovery, Vol. 11, pp. 195-212, Sep 2005. DOI: https://doi.org/10.1007/s10618-005-0007-5\n",
    "#### [4] Josep Domingo-Ferrer and Vicenç Torra, \"Disclosure risk assessment in statistical data protection\", Journal of Computational and Applied Mathematics, Vol. 164, pp. 285-293, Mar 2004. DOI: https://doi.org/10.1016/S0377-0427(03)00643-5\n",
    "### Adult data set (45222 records)\n",
    "#### k-anonymity via microaggregation (k=10) is much faster (~6 sec.) than via MDAV at the cost of utility (SSE = ~0.93)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "from anonymization.src.entities.dataset_DataFrame import Dataset_DataFrame\n",
    "from anonymization.src.algorithms.anonymization_scheme import Anonymization_scheme\n",
    "from anonymization.src.algorithms.k_anonymity import K_anonymity\n",
    "from anonymization.src.algorithms.microaggregation import Microaggregation\n",
    "from anonymization.src.utils import utils"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Following, it is indicated the path to the csv file containing the data set and the path to the xml file describing the attributes in the data set. Inside the xml file, there is a detailed descritpion about how to fill this xml file in order to properly configure the different attribute types in the data set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "path_csv = \"./anonymization/input_datasets/adult.csv\"\n",
    "path_settings = \"./anonymization/input_datasets/metadata_adult.xml\"\n",
    "data_frame = utils.read_dataframe_from_csv(path_csv)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.1 The data set is loaded from a DataFrame passed as parameter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading dataset\n",
      "Dataset loaded: ./anonymization/input_datasets/adult.csv\n",
      "Records loaded: 45222\n",
      "Dataset: ./anonymization/input_datasets/adult.csv\n",
      "Dataset head:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>education</th>\n",
       "      <th>marital-status</th>\n",
       "      <th>occupation</th>\n",
       "      <th>relationship</th>\n",
       "      <th>race</th>\n",
       "      <th>sex</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>native-country</th>\n",
       "      <th>prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>25</td>\n",
       "      <td>Private</td>\n",
       "      <td>11th</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Machine-op-inspct</td>\n",
       "      <td>Own-child</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>38</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Farming-fishing</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>50</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>28</td>\n",
       "      <td>Local-gov</td>\n",
       "      <td>Assoc-acdm</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Protective-serv</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>44</td>\n",
       "      <td>Private</td>\n",
       "      <td>Some-college</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Machine-op-inspct</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>34</td>\n",
       "      <td>Private</td>\n",
       "      <td>10th</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Other-service</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>30</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age  workclass     education      marital-status         occupation  \\\n",
       "0   25    Private          11th       Never-married  Machine-op-inspct   \n",
       "1   38    Private       HS-grad  Married-civ-spouse    Farming-fishing   \n",
       "2   28  Local-gov    Assoc-acdm  Married-civ-spouse    Protective-serv   \n",
       "3   44    Private  Some-college  Married-civ-spouse  Machine-op-inspct   \n",
       "4   34    Private          10th       Never-married      Other-service   \n",
       "\n",
       "    relationship   race   sex  hours-per-week native-country prediction  \n",
       "0      Own-child  Black  Male              40  United-States      <=50K  \n",
       "1        Husband  White  Male              50  United-States      <=50K  \n",
       "2        Husband  White  Male              40  United-States       >50K  \n",
       "3        Husband  Black  Male              40  United-States       >50K  \n",
       "4  Not-in-family  White  Male              30  United-States      <=50K  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Dataset description:\n",
      "Data set: ./anonymization/input_datasets/adult.csv\n",
      "Records: 45222\n",
      "Attributes:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Attribute_type</th>\n",
       "      <th>Sensitivity_type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>age</td>\n",
       "      <td>numerical_discrete</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>workclass</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>education</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>marital-status</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>occupation</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>relationship</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>race</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>sex</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>hours-per-week</td>\n",
       "      <td>numerical_discrete</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>native-country</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>prediction</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>confidential</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Name      Attribute_type  Sensitivity_type\n",
       "0              age  numerical_discrete  quasi_identifier\n",
       "1        workclass   plain_categorical  quasi_identifier\n",
       "2        education   plain_categorical  quasi_identifier\n",
       "3   marital-status   plain_categorical  quasi_identifier\n",
       "4       occupation   plain_categorical  quasi_identifier\n",
       "5     relationship   plain_categorical  quasi_identifier\n",
       "6             race   plain_categorical  quasi_identifier\n",
       "7              sex   plain_categorical  quasi_identifier\n",
       "8   hours-per-week  numerical_discrete  quasi_identifier\n",
       "9   native-country   plain_categorical  quasi_identifier\n",
       "10      prediction   plain_categorical      confidential"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "dataset = Dataset_DataFrame(data_frame, path_settings)\n",
    "dataset.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.2 The data set is anonymized, in this case, applying k_anonymity via microaggregation with a privacy level of k=10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Anonymizing k-Anonymity, k = 10 via Microaggregation\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8475151c094d4f54a7eaa11590194af8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:18\n"
     ]
    }
   ],
   "source": [
    "k = 10\n",
    "anonymization_scheme = K_anonymity(dataset, k)\n",
    "algorithm = Microaggregation()\n",
    "anonymization_scheme.calculate_anonymization(algorithm)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.3 Information loss metrics are calculated by comparing original and anonymized data set  \n",
    "Metrics calculated: Sum of Square Error (SSE) and, for each attribute, mean and variance\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Calculating information loss metrics\n",
      "\n",
      "Information loss metrics:\n",
      "SSE: 0.978\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Original mean</th>\n",
       "      <th>Anonymized mean</th>\n",
       "      <th>Original variance</th>\n",
       "      <th>Anonymized variance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>age</td>\n",
       "      <td>38.5479</td>\n",
       "      <td>38.5437</td>\n",
       "      <td>174.912587</td>\n",
       "      <td>115.331874</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>workclass</td>\n",
       "      <td>Private</td>\n",
       "      <td>Private</td>\n",
       "      <td>0.263478</td>\n",
       "      <td>0.061298</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>education</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>0.673102</td>\n",
       "      <td>0.523241</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>marital-status</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>0.534408</td>\n",
       "      <td>0.479457</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>occupation</td>\n",
       "      <td>Craft-repair</td>\n",
       "      <td>Craft-repair</td>\n",
       "      <td>0.866879</td>\n",
       "      <td>0.814913</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>relationship</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Husband</td>\n",
       "      <td>0.587236</td>\n",
       "      <td>0.519261</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>race</td>\n",
       "      <td>White</td>\n",
       "      <td>White</td>\n",
       "      <td>0.139733</td>\n",
       "      <td>0.024546</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>sex</td>\n",
       "      <td>Male</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.324952</td>\n",
       "      <td>0.296139</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>hours-per-week</td>\n",
       "      <td>40.938</td>\n",
       "      <td>40.9329</td>\n",
       "      <td>144.180908</td>\n",
       "      <td>97.457167</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>native-country</td>\n",
       "      <td>United-States</td>\n",
       "      <td>United-States</td>\n",
       "      <td>0.086905</td>\n",
       "      <td>0.001106</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>prediction</td>\n",
       "      <td>&lt;=50K</td>\n",
       "      <td>&lt;=50K</td>\n",
       "      <td>0.247844</td>\n",
       "      <td>0.247844</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Name       Original mean     Anonymized mean  Original variance  \\\n",
       "0              age             38.5479             38.5437         174.912587   \n",
       "1        workclass             Private             Private           0.263478   \n",
       "2        education             HS-grad             HS-grad           0.673102   \n",
       "3   marital-status  Married-civ-spouse  Married-civ-spouse           0.534408   \n",
       "4       occupation        Craft-repair        Craft-repair           0.866879   \n",
       "5     relationship             Husband             Husband           0.587236   \n",
       "6             race               White               White           0.139733   \n",
       "7              sex                Male                Male           0.324952   \n",
       "8   hours-per-week              40.938             40.9329         144.180908   \n",
       "9   native-country       United-States       United-States           0.086905   \n",
       "10      prediction               <=50K               <=50K           0.247844   \n",
       "\n",
       "    Anonymized variance  \n",
       "0            115.331874  \n",
       "1              0.061298  \n",
       "2              0.523241  \n",
       "3              0.479457  \n",
       "4              0.814913  \n",
       "5              0.519261  \n",
       "6              0.024546  \n",
       "7              0.296139  \n",
       "8             97.457167  \n",
       "9              0.001106  \n",
       "10             0.247844  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "information_loss = Anonymization_scheme.calculate_information_loss(dataset, anonymization_scheme.anonymized_dataset)\n",
    "information_loss.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.4 Disclosure risk is calculated via record linkage between anonymized and original data set  \n",
    "The disclosure risk estimates the percentage of anonymized records that correct match with the original ones"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ecbd45d238544896b23365fd142cf174",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Disclosure risk: 627.753 (1.39%)\n"
     ]
    }
   ],
   "source": [
    "disclosure_risk = Anonymization_scheme.calculate_fast_record_linkage(dataset, anonymization_scheme.anonymized_dataset)\n",
    "disclosure_risk.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.5 The anonymized data set can be saved as a csv formated file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Dataset saved: ./anonymization/output_datasets/adult_anom.csv'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "anonymization_scheme.save_anonymized_dataset(\"./anonymization/output_datasets/adult_anom.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 3.6 Also, the anonymized data set can be converted to DataFrame (to use as input data for other libraries)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>education</th>\n",
       "      <th>marital-status</th>\n",
       "      <th>occupation</th>\n",
       "      <th>relationship</th>\n",
       "      <th>race</th>\n",
       "      <th>sex</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>native-country</th>\n",
       "      <th>prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>36</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Transport-moving</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>56</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>37</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Sales</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>36</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>28</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Machine-op-inspct</td>\n",
       "      <td>Own-child</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>38</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>38</td>\n",
       "      <td>Private</td>\n",
       "      <td>Bachelors</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Exec-managerial</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>47</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>34</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Other-service</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>30</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  age workclass  education      marital-status         occupation  \\\n",
       "0  36   Private    HS-grad  Married-civ-spouse   Transport-moving   \n",
       "1  37   Private    HS-grad  Married-civ-spouse              Sales   \n",
       "2  28   Private    HS-grad       Never-married  Machine-op-inspct   \n",
       "3  38   Private  Bachelors  Married-civ-spouse    Exec-managerial   \n",
       "4  34   Private    HS-grad       Never-married      Other-service   \n",
       "\n",
       "    relationship   race   sex hours-per-week native-country prediction  \n",
       "0        Husband  White  Male             56  United-States      <=50K  \n",
       "1        Husband  White  Male             36  United-States      <=50K  \n",
       "2      Own-child  White  Male             38  United-States       >50K  \n",
       "3        Husband  White  Male             47  United-States       >50K  \n",
       "4  Not-in-family  White  Male             30  United-States      <=50K  "
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_anonymized = anonymization_scheme.anonymized_dataset_to_dataframe()\n",
    "df_anonymized.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "========================================================================================================================\n",
    "## 4. k-t-closeness via Microaggregation\n",
    "### References:\n",
    "#### [1] Josep Domingo-Ferrer and Vicenç Torra, \"Ordinal, continuous and heterogeneous k-anonymity through microaggregation\", Data Mining and Knowledge Discovery, Vol. 11, pp. 195-212, Sep 2005. DOI: https://doi.org/10.1007/s10618-005-0007-5\n",
    "#### [2] Jordi Soria-Comas, Josep Domingo-Ferrer, David Sánchez and Sergio Martínez, \"t-Closeness through microaggregation: strict privacy with enhanced utility preservation\", IEEE Transactions on Knowledge and Data Engineering, Vol. 27, no. 11, pp. 3098-3110, Oct 2015. DOI: https://doi.org/10.1109/TKDE.2015.2435777\n",
    "#### [4] Josep Domingo-Ferrer and Vicenç Torra, \"Disclosure risk assessment in statistical data protection\", Journal of Computational and Applied Mathematics, Vol. 164, pp. 285-293, Mar 2004. DOI: https://doi.org/10.1016/S0377-0427(03)00643-5\n",
    "### Adult data set (45222 records)\n",
    "#### k-t-closeness prevents attribute disclure"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "from anonymization.src.entities.dataset_DataFrame import Dataset_DataFrame\n",
    "from anonymization.src.algorithms.anonymization_scheme import Anonymization_scheme\n",
    "from anonymization.src.algorithms.t_closeness import T_closeness\n",
    "from anonymization.src.algorithms.microaggregation import Microaggregation\n",
    "from anonymization.src.utils import utils"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Following, it is indicated the path to the csv file containing the data set and the path to the xml file describing the attributes in the data set. Inside the xml file, there is a detailed descritpion about how to fill this xml file in order to properly configure the different attribute types in the data set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "path_csv = \"./anonymization/input_datasets/adult.csv\"\n",
    "path_settings = \"./anonymization/input_datasets/metadata_adult.xml\"\n",
    "data_frame = utils.read_dataframe_from_csv(path_csv)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4.1 The data set is loaded from a DataFrame passed as parameter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading dataset\n",
      "Dataset loaded: ./anonymization/input_datasets/adult.csv\n",
      "Records loaded: 45222\n",
      "Dataset: ./anonymization/input_datasets/adult.csv\n",
      "Dataset head:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>education</th>\n",
       "      <th>marital-status</th>\n",
       "      <th>occupation</th>\n",
       "      <th>relationship</th>\n",
       "      <th>race</th>\n",
       "      <th>sex</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>native-country</th>\n",
       "      <th>prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>25</td>\n",
       "      <td>Private</td>\n",
       "      <td>11th</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Machine-op-inspct</td>\n",
       "      <td>Own-child</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>38</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Farming-fishing</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>50</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>28</td>\n",
       "      <td>Local-gov</td>\n",
       "      <td>Assoc-acdm</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Protective-serv</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>44</td>\n",
       "      <td>Private</td>\n",
       "      <td>Some-college</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Machine-op-inspct</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>34</td>\n",
       "      <td>Private</td>\n",
       "      <td>10th</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Other-service</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>30</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age  workclass     education      marital-status         occupation  \\\n",
       "0   25    Private          11th       Never-married  Machine-op-inspct   \n",
       "1   38    Private       HS-grad  Married-civ-spouse    Farming-fishing   \n",
       "2   28  Local-gov    Assoc-acdm  Married-civ-spouse    Protective-serv   \n",
       "3   44    Private  Some-college  Married-civ-spouse  Machine-op-inspct   \n",
       "4   34    Private          10th       Never-married      Other-service   \n",
       "\n",
       "    relationship   race   sex  hours-per-week native-country prediction  \n",
       "0      Own-child  Black  Male              40  United-States      <=50K  \n",
       "1        Husband  White  Male              50  United-States      <=50K  \n",
       "2        Husband  White  Male              40  United-States       >50K  \n",
       "3        Husband  Black  Male              40  United-States       >50K  \n",
       "4  Not-in-family  White  Male              30  United-States      <=50K  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Dataset description:\n",
      "Data set: ./anonymization/input_datasets/adult.csv\n",
      "Records: 45222\n",
      "Attributes:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Attribute_type</th>\n",
       "      <th>Sensitivity_type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>age</td>\n",
       "      <td>numerical_discrete</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>workclass</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>education</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>marital-status</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>occupation</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>relationship</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>race</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>sex</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>hours-per-week</td>\n",
       "      <td>numerical_discrete</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>native-country</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>prediction</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>confidential</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Name      Attribute_type  Sensitivity_type\n",
       "0              age  numerical_discrete  quasi_identifier\n",
       "1        workclass   plain_categorical  quasi_identifier\n",
       "2        education   plain_categorical  quasi_identifier\n",
       "3   marital-status   plain_categorical  quasi_identifier\n",
       "4       occupation   plain_categorical  quasi_identifier\n",
       "5     relationship   plain_categorical  quasi_identifier\n",
       "6             race   plain_categorical  quasi_identifier\n",
       "7              sex   plain_categorical  quasi_identifier\n",
       "8   hours-per-week  numerical_discrete  quasi_identifier\n",
       "9   native-country   plain_categorical  quasi_identifier\n",
       "10      prediction   plain_categorical      confidential"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "dataset = Dataset_DataFrame(data_frame, path_settings)\n",
    "dataset.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4.2 The data set is anonymized, in this case, applying t-k-closeness via microaggregation with a privacy level of k=10 and t=0.25"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Anonymizing k-t-Closeness, k = 10, t = 0.25 via Microaggregation\n",
      "Sorting by confidential attribute: prediction\n",
      "Creating k subsets (10)\n",
      "Sorting each subset by quasi-identifiers\n",
      "Creating clusters\n",
      "Anonymizing\n",
      "Anonymization runtime: 0:00:13\n"
     ]
    }
   ],
   "source": [
    "k = 10\n",
    "t = 0.25\n",
    "anonymization_scheme = T_closeness(dataset, k, t)\n",
    "algorithm = Microaggregation()\n",
    "anonymization_scheme.calculate_anonymization(algorithm)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4.3 Information loss metrics are calculated by comparing original and anonymized data set  \n",
    "Metrics calculated: Sum of Square Error (SSE) and, for each attribute, mean and variance\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Calculating information loss metrics\n",
      "\n",
      "Information loss metrics:\n",
      "SSE: 1.106\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Original mean</th>\n",
       "      <th>Anonymized mean</th>\n",
       "      <th>Original variance</th>\n",
       "      <th>Anonymized variance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>age</td>\n",
       "      <td>38.5479</td>\n",
       "      <td>38.542</td>\n",
       "      <td>174.912587</td>\n",
       "      <td>44.272367</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>workclass</td>\n",
       "      <td>Private</td>\n",
       "      <td>Private</td>\n",
       "      <td>0.263478</td>\n",
       "      <td>0.032528</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>education</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>0.673102</td>\n",
       "      <td>0.477422</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>marital-status</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>0.534408</td>\n",
       "      <td>0.439388</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>occupation</td>\n",
       "      <td>Craft-repair</td>\n",
       "      <td>Craft-repair</td>\n",
       "      <td>0.866879</td>\n",
       "      <td>0.809363</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>relationship</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Husband</td>\n",
       "      <td>0.587236</td>\n",
       "      <td>0.469705</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>race</td>\n",
       "      <td>White</td>\n",
       "      <td>White</td>\n",
       "      <td>0.139733</td>\n",
       "      <td>0.005528</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>sex</td>\n",
       "      <td>Male</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.324952</td>\n",
       "      <td>0.245920</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>hours-per-week</td>\n",
       "      <td>40.938</td>\n",
       "      <td>40.9433</td>\n",
       "      <td>144.180908</td>\n",
       "      <td>62.766950</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>native-country</td>\n",
       "      <td>United-States</td>\n",
       "      <td>United-States</td>\n",
       "      <td>0.086905</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>prediction</td>\n",
       "      <td>&lt;=50K</td>\n",
       "      <td>&lt;=50K</td>\n",
       "      <td>0.247844</td>\n",
       "      <td>0.247844</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Name       Original mean     Anonymized mean  Original variance  \\\n",
       "0              age             38.5479              38.542         174.912587   \n",
       "1        workclass             Private             Private           0.263478   \n",
       "2        education             HS-grad             HS-grad           0.673102   \n",
       "3   marital-status  Married-civ-spouse  Married-civ-spouse           0.534408   \n",
       "4       occupation        Craft-repair        Craft-repair           0.866879   \n",
       "5     relationship             Husband             Husband           0.587236   \n",
       "6             race               White               White           0.139733   \n",
       "7              sex                Male                Male           0.324952   \n",
       "8   hours-per-week              40.938             40.9433         144.180908   \n",
       "9   native-country       United-States       United-States           0.086905   \n",
       "10      prediction               <=50K               <=50K           0.247844   \n",
       "\n",
       "    Anonymized variance  \n",
       "0             44.272367  \n",
       "1              0.032528  \n",
       "2              0.477422  \n",
       "3              0.439388  \n",
       "4              0.809363  \n",
       "5              0.469705  \n",
       "6              0.005528  \n",
       "7              0.245920  \n",
       "8             62.766950  \n",
       "9              0.000000  \n",
       "10             0.247844  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "information_loss = Anonymization_scheme.calculate_information_loss(dataset, anonymization_scheme.anonymized_dataset)\n",
    "information_loss.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4.4 Disclosure risk is calculated via record linkage between anonymized and original data set  \n",
    "The disclosure risk estimates the percentage of anonymized records that correct match with the original ones"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1bd2ca0bfc43444193f4c891af2f3d10",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Disclosure risk: 31.479 (0.07%)\n"
     ]
    }
   ],
   "source": [
    "disclosure_risk = Anonymization_scheme.calculate_fast_record_linkage(dataset, anonymization_scheme.anonymized_dataset)\n",
    "disclosure_risk.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4.5 The anonymized data set can be saved to a csv formated file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Dataset saved: ./anonymization/output_datasets/adult_anom.csv'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "anonymization_scheme.save_anonymized_dataset(\"./anonymization/output_datasets/adult_anom.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 4.6 Also, the anonymized data set can be converted to DataFrame (to use as input data for other libraries)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>education</th>\n",
       "      <th>marital-status</th>\n",
       "      <th>occupation</th>\n",
       "      <th>relationship</th>\n",
       "      <th>race</th>\n",
       "      <th>sex</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>native-country</th>\n",
       "      <th>prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>33</td>\n",
       "      <td>Private</td>\n",
       "      <td>Some-college</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Machine-op-inspct</td>\n",
       "      <td>Own-child</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>42</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>38</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Exec-managerial</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>29</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>38</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Craft-repair</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>41</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>31</td>\n",
       "      <td>Private</td>\n",
       "      <td>Some-college</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Sales</td>\n",
       "      <td>Own-child</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>38</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>44</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Other-service</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>35</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  age workclass     education      marital-status         occupation  \\\n",
       "0  33   Private  Some-college       Never-married  Machine-op-inspct   \n",
       "1  38   Private       HS-grad  Married-civ-spouse    Exec-managerial   \n",
       "2  38   Private       HS-grad  Married-civ-spouse       Craft-repair   \n",
       "3  31   Private  Some-college       Never-married              Sales   \n",
       "4  44   Private       HS-grad  Married-civ-spouse      Other-service   \n",
       "\n",
       "  relationship   race   sex hours-per-week native-country prediction  \n",
       "0    Own-child  White  Male             42  United-States      <=50K  \n",
       "1      Husband  White  Male             29  United-States      <=50K  \n",
       "2      Husband  White  Male             41  United-States       >50K  \n",
       "3    Own-child  White  Male             38  United-States       >50K  \n",
       "4      Husband  White  Male             35  United-States      <=50K  "
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_anonymized = anonymization_scheme.anonymized_dataset_to_dataframe()\n",
    "df_anonymized.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "========================================================================================================================\n",
    "## 5. Differential privacy via MDAV\n",
    "### References:\n",
    "#### [1] Josep Domingo-Ferrer and Vicenç Torra, \"Ordinal, continuous and heterogeneous k-anonymity through microaggregation\", Data Mining and Knowledge Discovery, Vol. 11, pp. 195-212, Sep 2005. DOI: https://doi.org/10.1007/s10618-005-0007-5\n",
    "#### [3] Jordi Soria-Comas, Josep Domingo-Ferrer, David Sánchez and Sergio Martínez, \"Enhancing data utility in differential privacy via microaggregation-based k-anonymity\", The VLDB Journal, Vol. 23, no. 5, pp. 771-794, Sep 2014. DOI: https://doi.org/10.1007/s00778-014-0351-4\n",
    "#### [4] Josep Domingo-Ferrer and Vicenç Torra, \"Disclosure risk assessment in statistical data protection\", Journal of Computational and Applied Mathematics, Vol. 164, pp. 285-293, Mar 2004. DOI: https://doi.org/10.1016/S0377-0427(03)00643-5\n",
    "### Adult data set (45222 records)\n",
    "#### Differential privacy offers privacy guarantees at the cost of the lost of the original data utility"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "from anonymization.src.entities.dataset_DataFrame import Dataset_DataFrame\n",
    "from anonymization.src.algorithms.anonymization_scheme import Anonymization_scheme\n",
    "from anonymization.src.algorithms.differential_privacy import Differential_privacy\n",
    "from anonymization.src.algorithms.microaggregation import Microaggregation\n",
    "from anonymization.src.utils import utils"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Following, it is indicated the path to the csv file containing the data set and the path to the xml file describing the attributes in the data set. Inside the xml file, there is a detailed descritpion about how to fill this xml file in order to properly configure the different attribute types in the data set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "path_csv = \"./anonymization/input_datasets/adult.csv\"\n",
    "path_settings = \"./anonymization/input_datasets/metadata_adult.xml\"\n",
    "data_frame = utils.read_dataframe_from_csv(path_csv)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5.1 The data set is loaded from a DataFrame passed as parameter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading dataset\n",
      "Dataset loaded: ./anonymization/input_datasets/adult.csv\n",
      "Records loaded: 45222\n",
      "Dataset: ./anonymization/input_datasets/adult.csv\n",
      "Dataset head:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>education</th>\n",
       "      <th>marital-status</th>\n",
       "      <th>occupation</th>\n",
       "      <th>relationship</th>\n",
       "      <th>race</th>\n",
       "      <th>sex</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>native-country</th>\n",
       "      <th>prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>25</td>\n",
       "      <td>Private</td>\n",
       "      <td>11th</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Machine-op-inspct</td>\n",
       "      <td>Own-child</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>38</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Farming-fishing</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>50</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>28</td>\n",
       "      <td>Local-gov</td>\n",
       "      <td>Assoc-acdm</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Protective-serv</td>\n",
       "      <td>Husband</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>44</td>\n",
       "      <td>Private</td>\n",
       "      <td>Some-college</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Machine-op-inspct</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Black</td>\n",
       "      <td>Male</td>\n",
       "      <td>40</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>34</td>\n",
       "      <td>Private</td>\n",
       "      <td>10th</td>\n",
       "      <td>Never-married</td>\n",
       "      <td>Other-service</td>\n",
       "      <td>Not-in-family</td>\n",
       "      <td>White</td>\n",
       "      <td>Male</td>\n",
       "      <td>30</td>\n",
       "      <td>United-States</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age  workclass     education      marital-status         occupation  \\\n",
       "0   25    Private          11th       Never-married  Machine-op-inspct   \n",
       "1   38    Private       HS-grad  Married-civ-spouse    Farming-fishing   \n",
       "2   28  Local-gov    Assoc-acdm  Married-civ-spouse    Protective-serv   \n",
       "3   44    Private  Some-college  Married-civ-spouse  Machine-op-inspct   \n",
       "4   34    Private          10th       Never-married      Other-service   \n",
       "\n",
       "    relationship   race   sex  hours-per-week native-country prediction  \n",
       "0      Own-child  Black  Male              40  United-States      <=50K  \n",
       "1        Husband  White  Male              50  United-States      <=50K  \n",
       "2        Husband  White  Male              40  United-States       >50K  \n",
       "3        Husband  Black  Male              40  United-States       >50K  \n",
       "4  Not-in-family  White  Male              30  United-States      <=50K  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Dataset description:\n",
      "Data set: ./anonymization/input_datasets/adult.csv\n",
      "Records: 45222\n",
      "Attributes:\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Attribute_type</th>\n",
       "      <th>Sensitivity_type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>age</td>\n",
       "      <td>numerical_discrete</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>workclass</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>education</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>marital-status</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>occupation</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>relationship</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>race</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>sex</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>hours-per-week</td>\n",
       "      <td>numerical_discrete</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>native-country</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>quasi_identifier</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>prediction</td>\n",
       "      <td>plain_categorical</td>\n",
       "      <td>confidential</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Name      Attribute_type  Sensitivity_type\n",
       "0              age  numerical_discrete  quasi_identifier\n",
       "1        workclass   plain_categorical  quasi_identifier\n",
       "2        education   plain_categorical  quasi_identifier\n",
       "3   marital-status   plain_categorical  quasi_identifier\n",
       "4       occupation   plain_categorical  quasi_identifier\n",
       "5     relationship   plain_categorical  quasi_identifier\n",
       "6             race   plain_categorical  quasi_identifier\n",
       "7              sex   plain_categorical  quasi_identifier\n",
       "8   hours-per-week  numerical_discrete  quasi_identifier\n",
       "9   native-country   plain_categorical  quasi_identifier\n",
       "10      prediction   plain_categorical      confidential"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "dataset = Dataset_DataFrame(data_frame, path_settings)\n",
    "dataset.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5.2 The data set is anonymized, in this case, applying differential privacy via microaggregation with a privacy level of k=10 and epsilon=1.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Anonymizing Differential_privacy, k = 10, epsilon = 1.0 via Microaggregation\n",
      "Anonymizing attribute: age (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "627c76de7fbb420eb9b0ae6b0588eccb",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: workclass (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b3b52c6eb3634f21bb61529ebda4d6ec",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: education (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "65443ebc41b1491a974284f40555919a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: marital-status (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0834a16f9dbc4b97a77d0d90e80c36af",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: occupation (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2a4d169d8a2345e59d3922085bfcab78",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: relationship (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8f83919004e541e38a2edf70d6bacd31",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: race (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bde9a2ebe7584474aed8b72d6c91216f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: sex (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d752ae013d2f4d5a9729ed663f1620bd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: hours-per-week (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c6e8c04e534142f1bd79bf56aa90fa88",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: native-country (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8150bb3017714b8089b6aba85b42a412",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:40\n"
     ]
    }
   ],
   "source": [
    "k = 10\n",
    "epsilon = 1.0\n",
    "anonymization_scheme = Differential_privacy(dataset, k, epsilon)\n",
    "algorithm = Microaggregation()\n",
    "anonymization_scheme.calculate_anonymization(algorithm)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5.3 Information loss metrics are calculated by comparing original and anonymized data set  \n",
    "Metrics calculated: Sum of Square Error (SSE) and, for each attribute, mean and variance\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Calculating information loss metrics\n",
      "\n",
      "Information loss metrics:\n",
      "SSE: 2.049\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Name</th>\n",
       "      <th>Original mean</th>\n",
       "      <th>Anonymized mean</th>\n",
       "      <th>Original variance</th>\n",
       "      <th>Anonymized variance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>age</td>\n",
       "      <td>38.5479</td>\n",
       "      <td>43.7774</td>\n",
       "      <td>174.912587</td>\n",
       "      <td>1783.103180</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>workclass</td>\n",
       "      <td>Private</td>\n",
       "      <td>Private</td>\n",
       "      <td>0.263478</td>\n",
       "      <td>0.499359</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>education</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>0.673102</td>\n",
       "      <td>0.559285</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>marital-status</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>0.534408</td>\n",
       "      <td>0.519216</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>occupation</td>\n",
       "      <td>Craft-repair</td>\n",
       "      <td>Craft-repair</td>\n",
       "      <td>0.866879</td>\n",
       "      <td>0.621158</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>relationship</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Husband</td>\n",
       "      <td>0.587236</td>\n",
       "      <td>0.546017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>race</td>\n",
       "      <td>White</td>\n",
       "      <td>White</td>\n",
       "      <td>0.139733</td>\n",
       "      <td>0.467295</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>sex</td>\n",
       "      <td>Male</td>\n",
       "      <td>Male</td>\n",
       "      <td>0.324952</td>\n",
       "      <td>0.436292</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>hours-per-week</td>\n",
       "      <td>40.938</td>\n",
       "      <td>46.7485</td>\n",
       "      <td>144.180908</td>\n",
       "      <td>1743.013401</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>native-country</td>\n",
       "      <td>United-States</td>\n",
       "      <td>United-States</td>\n",
       "      <td>0.086905</td>\n",
       "      <td>0.501968</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>prediction</td>\n",
       "      <td>&lt;=50K</td>\n",
       "      <td>&lt;=50K</td>\n",
       "      <td>0.247844</td>\n",
       "      <td>0.247844</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Name       Original mean     Anonymized mean  Original variance  \\\n",
       "0              age             38.5479             43.7774         174.912587   \n",
       "1        workclass             Private             Private           0.263478   \n",
       "2        education             HS-grad             HS-grad           0.673102   \n",
       "3   marital-status  Married-civ-spouse  Married-civ-spouse           0.534408   \n",
       "4       occupation        Craft-repair        Craft-repair           0.866879   \n",
       "5     relationship             Husband             Husband           0.587236   \n",
       "6             race               White               White           0.139733   \n",
       "7              sex                Male                Male           0.324952   \n",
       "8   hours-per-week              40.938             46.7485         144.180908   \n",
       "9   native-country       United-States       United-States           0.086905   \n",
       "10      prediction               <=50K               <=50K           0.247844   \n",
       "\n",
       "    Anonymized variance  \n",
       "0           1783.103180  \n",
       "1              0.499359  \n",
       "2              0.559285  \n",
       "3              0.519216  \n",
       "4              0.621158  \n",
       "5              0.546017  \n",
       "6              0.467295  \n",
       "7              0.436292  \n",
       "8           1743.013401  \n",
       "9              0.501968  \n",
       "10             0.247844  "
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "information_loss = Anonymization_scheme.calculate_information_loss(dataset, anonymization_scheme.anonymized_dataset)\n",
    "information_loss.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5.4 Disclosure risk is calculated via record linkage between anonymized and original data set  \n",
    "The disclosure risk estimates the percentage of anonymized records that correct match with the original ones"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "377779e9c1524ab28cc4b309d5ff0297",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Disclosure risk: 6.500 (0.01%)\n"
     ]
    }
   ],
   "source": [
    "disclosure_risk = Anonymization_scheme.calculate_fast_record_linkage(dataset, anonymization_scheme.anonymized_dataset)\n",
    "disclosure_risk.description()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5.5 The anonymized data set can be saved to a csv formated file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Dataset saved: ./anonymization/output_datasets/adult_anom.csv'"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "anonymization_scheme.save_anonymized_dataset(\"./anonymization/output_datasets/adult_anom.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 5.6 Also, the anonymized data set can be converted to DataFrame (to use as input data for other libraries)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>workclass</th>\n",
       "      <th>education</th>\n",
       "      <th>marital-status</th>\n",
       "      <th>occupation</th>\n",
       "      <th>relationship</th>\n",
       "      <th>race</th>\n",
       "      <th>sex</th>\n",
       "      <th>hours-per-week</th>\n",
       "      <th>native-country</th>\n",
       "      <th>prediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>81</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Armed-Forces</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Other</td>\n",
       "      <td>Male</td>\n",
       "      <td>1</td>\n",
       "      <td>Poland</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>44</td>\n",
       "      <td>Private</td>\n",
       "      <td>11th</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Armed-Forces</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Other</td>\n",
       "      <td>Male</td>\n",
       "      <td>99</td>\n",
       "      <td>Poland</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100</td>\n",
       "      <td>Without-pay</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Armed-Forces</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Other</td>\n",
       "      <td>Male</td>\n",
       "      <td>1</td>\n",
       "      <td>Poland</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>51</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Armed-Forces</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Other</td>\n",
       "      <td>Male</td>\n",
       "      <td>1</td>\n",
       "      <td>Poland</td>\n",
       "      <td>&gt;50K</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>Private</td>\n",
       "      <td>HS-grad</td>\n",
       "      <td>Married-civ-spouse</td>\n",
       "      <td>Armed-Forces</td>\n",
       "      <td>Husband</td>\n",
       "      <td>Other</td>\n",
       "      <td>Male</td>\n",
       "      <td>99</td>\n",
       "      <td>Poland</td>\n",
       "      <td>&lt;=50K</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age    workclass education      marital-status    occupation relationship  \\\n",
       "0   81      Private   HS-grad  Married-civ-spouse  Armed-Forces      Husband   \n",
       "1   44      Private      11th  Married-civ-spouse  Armed-Forces      Husband   \n",
       "2  100  Without-pay   HS-grad  Married-civ-spouse  Armed-Forces      Husband   \n",
       "3   51      Private   HS-grad  Married-civ-spouse  Armed-Forces      Husband   \n",
       "4    0      Private   HS-grad  Married-civ-spouse  Armed-Forces      Husband   \n",
       "\n",
       "    race   sex hours-per-week native-country prediction  \n",
       "0  Other  Male              1         Poland      <=50K  \n",
       "1  Other  Male             99         Poland      <=50K  \n",
       "2  Other  Male              1         Poland       >50K  \n",
       "3  Other  Male              1         Poland       >50K  \n",
       "4  Other  Male             99         Poland      <=50K  "
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_anonymized = anonymization_scheme.anonymized_dataset_to_dataframe()\n",
    "df_anonymized.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "scrolled": true
   },
   "source": [
    "========================================================================================================================\n",
    "## 6. Comparative study, adult data set (45222 records)\n",
    "### It is compared the information loss, disclosure risk and runtime for different methods and levels of privacy\n",
    "### The privacy level 'k' is ranged from 10 to 500"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from anonymization.src.entities.dataset_CSV import Dataset_CSV\n",
    "from anonymization.src.entities.dataset_DataFrame import Dataset_DataFrame\n",
    "from anonymization.src.algorithms.anonymization_scheme import Anonymization_scheme\n",
    "from anonymization.src.algorithms.k_anonymity import K_anonymity\n",
    "from anonymization.src.algorithms.t_closeness import T_closeness\n",
    "from anonymization.src.algorithms.differential_privacy import Differential_privacy\n",
    "from anonymization.src.algorithms.mdav import Mdav\n",
    "from anonymization.src.algorithms.microaggregation import Microaggregation\n",
    "from anonymization.src.utils import utils\n",
    "import pickle\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "plt.rcParams['figure.figsize'] = (8, 5)\n",
    "plt.style.use('ggplot')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Following, it is indicated the path to the csv file containing the data set and the path to the xml file describing the attributes in the data set. Inside the xml file, there is a detailed descritpion about how to fill this xml file in order to properly configure the different attribute types in the data set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "# k_list = [10, 25, 50, 100, 250, 500, 1000, 2500, 5000, 10000, 25000]\n",
    "k_list = [10, 25, 50, 100, 250, 500]\n",
    "path_csv = \"./anonymization/input_datasets/adult.csv\"\n",
    "path_settings = \"./anonymization/input_datasets/metadata_adult.xml\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6.1 k-Anonymity via mdav"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading dataset\n",
      "Dataset loaded: ./anonymization/input_datasets/adult.csv\n",
      "Records loaded: 45222\n",
      "Anonymizing k-Anonymity, k = 10 via MDAV\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "de8d1df2c548459b919f8277972286a1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 1:04:52\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "dcf7b4d5152e4c0c81516a21c1d25378",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-Anonymity, k = 25 via MDAV\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "26faae7c07764b25b7525cf87260ce6e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:25:29\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6234a716f18d4743b7923c9a112baefd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-Anonymity, k = 50 via MDAV\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6096fff5f0a648ca8c94de49e53d48e6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:12:45\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e891afb9302b4f1b9112e3f7158a4ede",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-Anonymity, k = 100 via MDAV\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "20f92a1529f843c69bbb4a7525852221",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:06:26\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "fcb9ad548c1b4eb0ac57b8a731a9d195",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-Anonymity, k = 250 via MDAV\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3b39fdf3b45a4565a9d91612f27e6930",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:02:40\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6271939e676c45bb89ae4abe75c7378d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-Anonymity, k = 500 via MDAV\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "595d6e0911064ae7bd3cffa9f458121a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:01:24\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2cd52e5e444844c8865f684a3abd349a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "data_frame = utils.read_dataframe_from_csv(path_csv)\n",
    "dataset = Dataset_DataFrame(data_frame, path_settings)\n",
    "il_k_anom_mdav = []\n",
    "dr_k_anom_mdav = []\n",
    "run_k_anom_mdav = []\n",
    "for k in k_list:\n",
    "    anonymization_scheme = K_anonymity(dataset, k)\n",
    "    algorithm = Mdav()\n",
    "    anonymization_scheme.calculate_anonymization(algorithm)\n",
    "    run_k_anom_mdav.append(Anonymization_scheme.runtime)\n",
    "    il_k_anom_mdav.append(Anonymization_scheme.calculate_information_loss(dataset, anonymization_scheme.anonymized_dataset))\n",
    "    dr_k_anom_mdav.append(Anonymization_scheme.calculate_fast_record_linkage(dataset, anonymization_scheme.anonymized_dataset))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6.2 k-Anonymity via microaggregation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading dataset\n",
      "Dataset loaded: ./anonymization/input_datasets/adult.csv\n",
      "Records loaded: 45222\n",
      "Anonymizing k-Anonymity, k = 10 via Microaggregation\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bbb81ace941d42d38e7ff472390f5bb6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:18\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "28e8e4f79a834cdb9682c5f35d19f736",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-Anonymity, k = 25 via Microaggregation\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "af1c8b0a123a4f8ab859f8c3fc7178ea",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:15\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bf352f6a47aa48ce8f44c4d48126c0dd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-Anonymity, k = 50 via Microaggregation\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "497c40d4287f4ba5bb63cbc717ad248f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:16\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8ce0428d025148eb80bcfd9987c26313",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-Anonymity, k = 100 via Microaggregation\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "19e2669171cf4fe481b217735ebb96ba",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:15\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "14cf5c8602364860a580f3830d5faa33",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-Anonymity, k = 250 via Microaggregation\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "309d8accc166489794064f86d3565d4d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:16\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "dcb2d946f96c480e8fdfe1e449884c58",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-Anonymity, k = 500 via Microaggregation\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a06c279258364392be8c156f096835cd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:14\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a8d614277109407e9af195798d9574c8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "data_frame = utils.read_dataframe_from_csv(path_csv)\n",
    "dataset = Dataset_DataFrame(data_frame, path_settings)\n",
    "il_k_anom_micro = []\n",
    "dr_k_anom_micro = []\n",
    "run_k_anom_micro = []\n",
    "for k in k_list:\n",
    "    anonymization_scheme = K_anonymity(dataset, k)\n",
    "    algorithm = Microaggregation()\n",
    "    anonymization_scheme.calculate_anonymization(algorithm)\n",
    "    run_k_anom_micro.append(Anonymization_scheme.runtime)\n",
    "    il_k_anom_micro.append(Anonymization_scheme.calculate_information_loss(dataset, anonymization_scheme.anonymized_dataset))\n",
    "    dr_k_anom_micro.append(Anonymization_scheme.calculate_fast_record_linkage(dataset, anonymization_scheme.anonymized_dataset))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6.3 k-t-Closeness via microaggregation, t=0.25 (maximum attribute disclosure protection) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading dataset\n",
      "Dataset loaded: ./anonymization/input_datasets/adult.csv\n",
      "Records loaded: 45222\n",
      "Anonymizing k-t-Closeness, k = 10, t = 0.25 via Microaggregation\n",
      "Sorting by confidential attribute: prediction\n",
      "Creating k subsets (10)\n",
      "Sorting each subset by quasi-identifiers\n",
      "Creating clusters\n",
      "Anonymizing\n",
      "Anonymization runtime: 0:00:14\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ebd9c53ca66946d6944bc9af3fb7e4c4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-t-Closeness, k = 25, t = 0.25 via Microaggregation\n",
      "Sorting by confidential attribute: prediction\n",
      "Creating k subsets (25)\n",
      "Sorting each subset by quasi-identifiers\n",
      "Creating clusters\n",
      "Anonymizing\n",
      "Anonymization runtime: 0:00:14\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7f6e3e4d21134d76bd04ddc0e5fefa88",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-t-Closeness, k = 50, t = 0.25 via Microaggregation\n",
      "Sorting by confidential attribute: prediction\n",
      "Creating k subsets (50)\n",
      "Sorting each subset by quasi-identifiers\n",
      "Creating clusters\n",
      "Anonymizing\n",
      "Anonymization runtime: 0:00:14\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e430ae24ae9e4ea89cbc9b027cfe7f2e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-t-Closeness, k = 100, t = 0.25 via Microaggregation\n",
      "Sorting by confidential attribute: prediction\n",
      "Creating k subsets (100)\n",
      "Sorting each subset by quasi-identifiers\n",
      "Creating clusters\n",
      "Anonymizing\n",
      "Anonymization runtime: 0:00:14\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6fb4b9a69de748228e0717ecf5b21e34",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-t-Closeness, k = 250, t = 0.25 via Microaggregation\n",
      "Sorting by confidential attribute: prediction\n",
      "Creating k subsets (251)\n",
      "Sorting each subset by quasi-identifiers\n",
      "Creating clusters\n",
      "Anonymizing\n",
      "Anonymization runtime: 0:00:14\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "85b768dc6ddc4a988e8d1d20e6421948",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing k-t-Closeness, k = 500, t = 0.25 via Microaggregation\n",
      "Sorting by confidential attribute: prediction\n",
      "Creating k subsets (502)\n",
      "Sorting each subset by quasi-identifiers\n",
      "Creating clusters\n",
      "Anonymizing\n",
      "Anonymization runtime: 0:00:14\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "03a028992fe944929dbef3eed8d52a9a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "data_frame = utils.read_dataframe_from_csv(path_csv)\n",
    "dataset = Dataset_DataFrame(data_frame, path_settings)\n",
    "il_t_clos = []\n",
    "dr_t_clos = []\n",
    "run_t_clos = []\n",
    "t = 0.25\n",
    "for k in k_list:\n",
    "    anonymization_scheme = T_closeness(dataset, k, t)\n",
    "    algorithm = Microaggregation()\n",
    "    anonymization_scheme.calculate_anonymization(algorithm)\n",
    "    run_t_clos.append(Anonymization_scheme.runtime)\n",
    "    il_t_clos.append(anonymization_scheme.calculate_information_loss(dataset, anonymization_scheme.anonymized_dataset))\n",
    "    dr_t_clos.append(anonymization_scheme.calculate_fast_record_linkage(dataset, anonymization_scheme.anonymized_dataset))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6.4 Differential privacy via individual ranking through mdav, epsilon=1.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading dataset\n",
      "Dataset loaded: ./anonymization/input_datasets/adult.csv\n",
      "Records loaded: 45222\n",
      "Anonymizing Differential_privacy, k = 10, epsilon = 1.0 via MDAV\n",
      "Anonymizing attribute: age (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2fd8ab7345f145f4afd83768cbb1b812",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: workclass (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "133ded5e227d442c913fcc46f3782606",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: education (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7f2ad01793da4eb190dcb7308e0640fe",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: marital-status (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ea7e75413dfd4699a83fe90ff5d8ecbd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: occupation (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "92e8231a20ca4cfa9caecebf58cf07bd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: relationship (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9918da62cd5544d4bcb447b009289c0e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: race (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "cee67a1d6f9644d590787e484036662c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: sex (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b560f7e96aeb4bdd864bb88f52384cb6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: hours-per-week (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2c33a24b63824bb1bceaca4b4d5d8d46",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: native-country (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4c5548ffa2bd4d56b1c3082d683b9c5a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 1:46:21\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "14b10774c8c240a49fed53f77e81bfcd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing Differential_privacy, k = 25, epsilon = 1.0 via MDAV\n",
      "Anonymizing attribute: age (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "83413ee828ca44ca95e7539d5c343626",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: workclass (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "aac398079dc94b96b8fa35d96164953d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: education (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bf6188c11ac44fee83e83a63111fe7f0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: marital-status (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3f0c706f30f640a88ce69667218986db",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: occupation (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "beb37b56eb9f43b5b0853e5304be43c2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: relationship (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "131bab8598c84edd987295b7eabeb82e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: race (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9edad162e8654046ad95a51814281692",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: sex (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "35a5b884be664cb8ae5294c4f11b2c27",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: hours-per-week (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e2f90a06c4744735a692f58bbe5ec985",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: native-country (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "be3100c5f96d432088fd0943f2eeb3f6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:42:55\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9e0c67d35ad94377b099b50c343e00bd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing Differential_privacy, k = 50, epsilon = 1.0 via MDAV\n",
      "Anonymizing attribute: age (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0c865c6bb02448c0b0379bf2c6774150",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: workclass (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b014057469ed48f4892bf36d90d4f39a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: education (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0032b81f1d904cd18442ee6d9bfbfeb4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: marital-status (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0e4de99ab6b2493795a66874bb43655d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: occupation (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c103950f0e324c24ab5148636e23fc96",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: relationship (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e94dd67508f440c3a991c41c08140fa2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: race (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "29259b683b914f7999b88a1209c832de",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: sex (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a788f4b481134823a250d6b4f2eced08",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: hours-per-week (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a638fee41e554104a91b93aa38a4da14",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: native-country (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "81d87e93c37343b8a604cafcfebd7f25",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:21:28\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "37a1ae83ce004c74b10cefc8f7eb3ea0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing Differential_privacy, k = 100, epsilon = 1.0 via MDAV\n",
      "Anonymizing attribute: age (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d958d58f7b6245e991ec1e66a9f158e7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: workclass (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f51129c2ca504541991b392aa4bd6174",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: education (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "36585b6a8e8840c08539d936f1c3bfc1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: marital-status (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2b5d71c437374978ab680a6f5823fdf4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: occupation (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "cb7e6a1f96b84cc68ffdb5fce46d2273",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: relationship (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0663badd5f674b7f80e27c8054b3fa33",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: race (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d02ab1582f194f658534695c0a935feb",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: sex (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e7b132d68e3141d98f90234510de5169",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: hours-per-week (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5d2580d6b06642dc8d5d75bcdf96a915",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: native-country (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e0c06a5c899845e7869093d72d05e630",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:10:56\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8f4ca8b028c04607ab535f7db81fad27",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing Differential_privacy, k = 250, epsilon = 1.0 via MDAV\n",
      "Anonymizing attribute: age (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "71883b85bfbf441b8aabcd98a9fae3cb",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: workclass (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9cff01ce653442b79bba79bdd5434532",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: education (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "eba9768847cb4500abbef5bbbc189f6f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: marital-status (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bd012d9512d74cf2984a562960f1daa8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: occupation (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d3a3834c271d40fb9c269be028018ec3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: relationship (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7e150bd73e9b40819073b6de807827d6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: race (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2acc596c4dd348dda9e42f4d37856707",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: sex (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a155b718d4ca45b780a6e40b793b84f0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: hours-per-week (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b5998e13477841da892d9ec8968b0d35",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: native-country (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a60320a8c723443ea8d5e77476af24cc",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:04:33\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d667693166ed451aae094d6fb9579e8c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing Differential_privacy, k = 500, epsilon = 1.0 via MDAV\n",
      "Anonymizing attribute: age (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6da16b3e87d74989a4c3209a42f513dd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: workclass (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7335f671851d4b0789ed672df794e5e4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: education (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d4dd1b6de1a04ddfaf8dd6c13916a2fe",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: marital-status (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "69e78afb5292425bb761694561b1deb3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: occupation (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4af23bd55d254f2c9bae4c65001ea9ea",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: relationship (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ecc4cf4acd49490e86f6e9b52e033fe1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: race (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e925ad8cf5f44bbcbfab260d6e7387df",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: sex (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e5abb3b2fbc946d08bcd301093cb5639",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: hours-per-week (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a449aed00bf046518ec76a40194c0446",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: native-country (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c71421c26b604bde8a990823fcce5304",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:02:26\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d013c8b08ba1406f9812388d4bf892bd",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "data_frame = utils.read_dataframe_from_csv(path_csv)\n",
    "dataset = Dataset_DataFrame(data_frame, path_settings)\n",
    "il_dp_mdav = []\n",
    "dr_dp_mdav = []\n",
    "run_dp_mdav = []\n",
    "epsilon = 1.0\n",
    "for k in k_list:\n",
    "    anonymization_scheme = Differential_privacy(dataset, k, epsilon)\n",
    "    algorithm = Mdav()\n",
    "    anonymization_scheme.calculate_anonymization(algorithm)\n",
    "    run_dp_mdav.append(Anonymization_scheme.runtime)\n",
    "    il_dp_mdav.append(anonymization_scheme.calculate_information_loss(dataset, anonymization_scheme.anonymized_dataset))\n",
    "    dr_dp_mdav.append(anonymization_scheme.calculate_fast_record_linkage(dataset, anonymization_scheme.anonymized_dataset))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6.5 Differential privacy via individual ranking through microaggregation, epsilon=1.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading dataset\n",
      "Dataset loaded: ./anonymization/input_datasets/adult.csv\n",
      "Records loaded: 45222\n",
      "Anonymizing Differential_privacy, k = 10, epsilon = 1.0 via Microaggregation\n",
      "Anonymizing attribute: age (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e04b723a53c449b4b99eef1104b0fbd6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: workclass (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d0d4139356ff43fa858722dd3b9c1e5a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: education (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2863cae510eb41a6aa5da6b21cedd836",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: marital-status (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "38c5aefb7da44ba6822dcd2d502d5447",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: occupation (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a5aa28077b99492d88050b4014128067",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: relationship (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "20c71b42af2d498f85a0b3282a185e82",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: race (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "83f9f86bd1174d199a0171bfbd6afc20",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: sex (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "445e0c9fe0554cf89aeb7757165e8053",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: hours-per-week (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "30510abb9eff4d7883d4824066bec06c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: native-country (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8a01e0f526a64d84b78e0963b858e1e9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:50\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "eb9667735d3045c586f03d14e2a1f8e1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing Differential_privacy, k = 25, epsilon = 1.0 via Microaggregation\n",
      "Anonymizing attribute: age (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "dc6bab0addaf4ef58812e9b75778b428",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: workclass (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "cb2033cf1891439f98164dd5108669c8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: education (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5d8d8d73e66c478da7fbe2a009b1e754",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: marital-status (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ed187c9577cc4d09b28622a39bcad2c9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: occupation (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "70a2a4e4b78241f6b03cfc024c5a7740",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: relationship (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ec63d54325704f89aab3f9cc01d1bcad",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: race (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "334f1d4704d14fe9963ba890fd1f0e96",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: sex (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d447eaa22e0f4ac4a23af3cb752a4417",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: hours-per-week (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a9c8ac32e08b4c9980c27a92c7acb7f5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: native-country (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "268a0c3a5c824aaf9afc8ed1365d5ec3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:37\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4656987f247441ab86ba453b25faccef",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing Differential_privacy, k = 50, epsilon = 1.0 via Microaggregation\n",
      "Anonymizing attribute: age (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b1f17c4fafab4fa5ac3af50eecb2ff99",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: workclass (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "502624b1c796427e84218627b895f11e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: education (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "828ff20a2ebe41959060cad80d2bf3ac",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: marital-status (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6a7c3c357d4d40e198894b86269893d0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: occupation (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2936350ec3024970a764b5654017a7ce",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: relationship (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1eac236e162a483dbb19373af743dc8c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: race (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "634b4d16272e4296b826d315aa541295",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: sex (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3fd7f14b88a74428959776ae522b1a3c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: hours-per-week (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9c6ff4d7165744ca930f625f1885ac2f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: native-country (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3aec5503e1df49f48a4a3df7289a6809",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:33\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ef12f39f00474b0b84473e64bd68ee17",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing Differential_privacy, k = 100, epsilon = 1.0 via Microaggregation\n",
      "Anonymizing attribute: age (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a789afa6a9e64795a6d06f4006501259",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: workclass (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "cc426b2926d14abe8022b8747f8c5fca",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: education (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3979c9ca0fcd4047a91f5b817b93f9a9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: marital-status (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f32a062af5b34849ba109d5a310427b1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: occupation (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "783ccf643dbc4fd6bc1e294359cdca99",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: relationship (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8903f7bc0abb41a68f45573b2af02930",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: race (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "13429c36a7114202824549519f651510",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: sex (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4137fa8a527b4d6bb2ae868267feb155",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: hours-per-week (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "fb19b17f37a745258715d30f7e56c3f2",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: native-country (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "78dd4b82f46f4d0782297c2645215bd7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:31\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4bd611ded4de462dbb529ada348d7bb9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing Differential_privacy, k = 250, epsilon = 1.0 via Microaggregation\n",
      "Anonymizing attribute: age (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ac817c7118b84b29b0355e9092f35058",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: workclass (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ac1db326bfe141caaf979db2042ec86a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: education (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ce9234e69702466abb59661e6280ca23",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: marital-status (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "7bb4a99eb7754dcfad057800ba0b5f0c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: occupation (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a29ddcf996c948e4adf85924062e8d5a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: relationship (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "95971171b671488e9de2787ab066a522",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: race (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "130845c1f12f4ee793046a1dd5796258",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: sex (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "70338bdc70df452bad7d3763db962284",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: hours-per-week (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9ac059a32a5245118fd05f0866ce9d6a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: native-country (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e565764914e5433a9f26e82c660522e3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:29\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "eb6ff809e16d448bab19dc51b6774cab",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing Differential_privacy, k = 500, epsilon = 1.0 via Microaggregation\n",
      "Anonymizing attribute: age (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3268db415cb34aa49c449c0c52daa651",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: workclass (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "74a8c70ff6cb4be0b6bd87c392d2acb0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: education (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1b3af69f275a456c957da8fc916460e7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: marital-status (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "8d5a7e6a607448e996e8714c60a8e88c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: occupation (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a17e8d19015f4ce0862119873bdb1288",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: relationship (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5532ca59e68545f1a23dea3114b9dc1d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: race (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9ec9f784668f4192bcee020b58557de6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: sex (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "996ee9d4a7a74a43af00c740b15cb857",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: hours-per-week (numerical_discrete)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a5fdbe89cfc740a1bb5385b0ff6abe7e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymizing attribute: native-country (plain_categorical)\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1f4d02b52fca4bfabf5542703a1c74ae",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Anonymization runtime: 0:00:32\n",
      "Calculating information loss metrics\n",
      "Calculating fast record linkage (disclosure risk), window size = 452.22\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9f268c4b246c4e4d9fb9be9bcb2b49e5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=45222.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "data_frame = utils.read_dataframe_from_csv(path_csv)\n",
    "dataset = Dataset_DataFrame(data_frame, path_settings)\n",
    "il_dp_micro = []\n",
    "dr_dp_micro = []\n",
    "run_dp_micro = []\n",
    "epsilon = 1.0\n",
    "for k in k_list:\n",
    "    anonymization_scheme = Differential_privacy(dataset, k, epsilon)\n",
    "    algorithm = Microaggregation()\n",
    "    anonymization_scheme.calculate_anonymization(algorithm)\n",
    "    run_dp_micro.append(Anonymization_scheme.runtime)\n",
    "    il_dp_micro.append(anonymization_scheme.calculate_information_loss(dataset, anonymization_scheme.anonymized_dataset))\n",
    "    dr_dp_micro.append(anonymization_scheme.calculate_fast_record_linkage(dataset, anonymization_scheme.anonymized_dataset))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6.6 Comparative of methods\n",
    "#### Information loss (SSE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20,6))\n",
    "k_str = [str(k) for k in k_list]\n",
    "l = [x.SSE for x in il_k_anom_micro]\n",
    "plt.plot(k_str, l, label='k-Anonymity Microaggregation')\n",
    "l = [x.SSE for x in il_k_anom_mdav]\n",
    "plt.plot(k_str, l, label='k-Anonymity MDAV')\n",
    "l = [x.SSE for x in il_t_clos]\n",
    "plt.plot(k_str, l, label='k-t-Closeness')\n",
    "l = [x.SSE for x in il_dp_micro]\n",
    "plt.plot(k_str, l, label='Differential privacy microaggregation')\n",
    "l = [x.SSE for x in il_dp_mdav]\n",
    "plt.plot(k_str, l, label='Differential privacy MDAV')\n",
    "plt.legend(loc='best', shadow=True, fontsize='12')\n",
    "plt.xlabel('k')\n",
    "plt.ylabel('SSE')\n",
    "plt.title('Information loss')\n",
    "plt.xticks(k_str)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6.7 Comparative of methods\n",
    "#### Disclosure risk (% matching in record linkage)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20,6))\n",
    "k_str = [str(k) for k in k_list]\n",
    "l = [x.percen for x in dr_k_anom_micro]\n",
    "plt.plot(k_str, l, label='k-Anonymity microaggregation')\n",
    "l = [x.percen for x in dr_k_anom_mdav]\n",
    "plt.plot(k_str, l, label='k-Anonymity MDAV')\n",
    "l = [x.percen for x in dr_t_clos]\n",
    "plt.plot(k_str, l, label='k-t-Closeness')\n",
    "l = [x.percen for x in dr_dp_micro]\n",
    "plt.plot(k_str, l, label='Differential privacy microaggregation')\n",
    "l = [x.percen for x in dr_dp_mdav]\n",
    "plt.plot(k_str, l, label='Differential privacy MDAV')\n",
    "plt.legend(loc='best', shadow=True, fontsize='12')\n",
    "plt.xlabel('k')\n",
    "plt.ylabel('% matching in record linkage')\n",
    "plt.title('Disclosure risk')\n",
    "plt.xticks(k_str)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6.8 Comparative of methods\n",
    "#### Runtime (seconds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABJsAAAGHCAYAAAAX21NaAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAACjsElEQVR4nOzdd3SUZf7+8fczJb33BtKSEHqkCtIUKQLSFV2QooK6NmQVy9pw185iWQFRF11Rvi5FELEAAqKigBQrSBdSKZkUUmcy8/sjOj8jbRDIhHC9zuEc5qnXM5k7OB/vYrhcLhciIiIiIiIiIiJngcnbAUREREREREREpO5QsUlERERERERERM4aFZtEREREREREROSsUbFJRERERERERETOGhWbRERERERERETkrFGxSUREREREREREzhoVm0RERERqqTfeeAOLxeLtGCIiIiKnRcUmERERkRMYO3YshmFgGAZms5mkpCSuv/56MjMzz+p9MjIyMAyDNWvWVNt+zTXXnPV7iYiIiJxrKjaJiIiInETXrl3Jzs5m//79vPPOO2zZsoURI0bUyL39/f2JjY2tkXuJiIiInC0qNomIiIichI+PD3FxcSQmJtKtWzcmTJjAV199RWFh4XGHuf2xl9KaNWswDIMVK1bQrVs3AgICaNasGZ988on7nHr16gHQs2dPDMOgQYMGwLHD6H57vXr1alq2bIm/vz/du3cnKyuLtWvXkp6eTmBgIL169TqmR9SKFSvo0qUL/v7+JCYmMm7cOI4cOXIO3jERERG50KnYJCIiIuKhrKwsFixYgNlsxmw2n9a5f/vb33jggQf49ttvadeuHddccw35+fkAbN68GYCFCxeSnZ3Nxo0bT3gdp9PJY489xmuvvcaXX35JVlYW11xzDQ8//DAzZ87kiy++ICMjg7vvvtt9zqpVqxg0aBAjR47ku+++Y/Hixezbt48hQ4bgcrlO/40QEREROQnNOCkiIiJyEmvWrCEoKAin00lpaSkAkydPJjAw8LSu88gjj9C3b18AnnnmGd566y3Wr19Pnz59iI6OBiAiIoK4uLiTXsflcvH888/Tpk0bACZMmMC9997LN998Q9u2bQGYOHEi//znP93nTJ06lTvuuIPbb7/dve3NN9/koosu4ttvv3VfS0RERORsUM8mERERkZPo2LEjW7duZcOGDTz00EN06tSJxx9//LSv8/uCTlxcHGazmdzc3NO+jmEYtGzZstq1AFq1alVt25EjR6isrARg48aNPP/88wQFBbn/NGvWDICdO3eedgYRERGRk1HPJhEREZGT8Pf3p0mTJgC0aNGCHTt28Ne//pX//Oc/mEzH/n87u91+3Ov4+Pgcs83pdJ52HpPJVG0In2EYAFit1mO2/TZEzul0MmXKFEaPHn3M9U7Vk0pERETkdKnYJCIiInIaHn30UZo3b86tt95KTEwMlZWV5ObmuleN+23+pdPxWyHqt55IZ1u7du348ccf3UUzERERkXNJw+hERERETkPTpk0ZMGAA999/Px06dCA4OJj77ruPnTt38vHHHzN16tTTvmZUVBRBQUEsX76cnJwcbDbbWc08depUlixZwqRJk9i6dSu7d+/m448/5oYbbnDPQyUiIiJytqjYJCIiInKa7r33XlauXMmWLVuYN28eX3/9Na1ateLxxx/nmWeeOe3rmUwmXn75Zf73v/9Rr1490tPTz2renj17smrVKr7//nu6du1Kq1atmDRpEsHBwdWG34mIiIicDYZL692KiIiIiIiIiMhZop5NIiIiIiIiIiJy1qjYJCIiIiIiIiIiZ42KTSIiIiIiIiIictao2CQiIiIiIiIiImeNik0iIiIiIiIiInLWqNgkIiIiIiIiIiJnjcXbAWpKVlaWtyOcsaioKA4fPuztGCK1ntqKiGfUVkQ8o7Yi4hm1FRHP1KW2kpCQcNzt6tkkIiIiIiIiIiJnjYpNIiIiIiIiIiJy1qjYJCIiIiIiIiIiZ42KTSIiIiIiIiIictZcMBOEi4iIiIiISBWXy0VxcTEul+usXbO0tBSHw3HWridSV51PbcXlcuHn54ePj89pnadik4iIiIiIyAWmuLgYX19frFart6OISC3mcrkoKCjg4MGDJCUleXyehtGJiIiIiIhcYFwulwpNInJKhmEQFhZGWVkZP/30k8fnqdgkIiIiIiIiIiInZDab2bFjh8fHq9gkIiIiIiIiIiInVVlZ6fGxKjaJiIiIiIhIrdGxY0fWrl3r7Rg1JjMzk+Tk5NP6In+2rF+/nq5du9b4fc9nU6ZMYfr06d6OUeup2CQiIiIiIiLntbvuuov69euTk5Pj7SinLTExkZ07d2I2mwEYPnw477zzTo3cu2PHjnz++ec1cq/z0bvvvsvgwYOrbXv66aeZNGmSdwKdR1RsOl+4KuHwBjiLS5OKiIiIiIic70pKSvjwww8JDg7mvffe83acC4LD4fB2BDdv9AiTU1Ox6TzhV7QZ065XCT60uKrwJCIiIiIiUsft2rWLTp06sWTJkhMes2zZMkJCQpg0aRLz58+vtm/atGlMnDiRO+64g5SUFHr27Mm3337r3r9z506GDx9OWloaPXv2ZPny5e59d911Fw888ACjR48mJSWFAQMGsG/fPgAeeOABHnvssWr3GjNmDK+++ipQ1WNo5syZ9OrViyZNmjB58mQOHTrEqFGjSElJ4ZprriE/Px+AAwcOkJiYiMPh4KmnnmL9+vX8/e9/Jzk5mQcffPCU9/qjxMRE3njjDbp06UJKSgrPPPMM+/btY+DAgaSmpjJx4kQqKioAWLduHW3btnWfm5mZyY033kjLli1p3rw5Dz74IFDVw2fQoEE88sgjNG/enGnTplFYWMgdd9xBy5Yt6dChA88//zxOpxOAffv2MWLECJo3b06LFi247bbbKCgocN/n+++/p3fv3qSkpDBhwgRuvvlmnn76aff+GTNmkJ6ezsUXX8w777xDYmIie/fudf9c7rvvPkaPHk2TJk348ssvycnJ4aabbqJly5Z06tSJ119/3X2t0tJS7rzzTpo1a0b37t2ZMWNGtWf+97//TefOnUlJSaFHjx589NFH7s/G/fffz6ZNm0hOTiYtLc19/99nffvtt+nSpQvNmzdn7Nix1XrXJSYm8t///pcuXbrQrFkzHnjgAVwXSAcSi7cDiGfKgtsSbCnFP+sjTI4CCuOuxWXy9XYsERERERE5zzn/71VcB/ae03sY9RpiGnnTaZ3z/fffM378eJ544gmuuOKKEx43f/58Bg8ezKBBg5g6dSrff/89LVu2dO9fsWIFr776KtOnT+eZZ57hwQcf5IMPPsButzN27FiuueYa3nnnHTZs2MD48eP58MMPadKkCQBLlixh7ty5tGzZ0l1kmDlzJiNGjOCGG27goYcewmQykZeXxxdffMFzzz3nvu+yZcuYN28eDoeDPn368MMPPzBt2jSSk5MZNWoU//nPf7j77rurPct9993HN998w9ChQ7nuuusA2LJlyynv9Udr1qzh448/Jisri759+/LNN9/w73//m/DwcK666ioWL17M1VdfXe2cyspKxowZQ5cuXVi/fj0mk4nvvvvOvX/Lli0MGjSIb7/9FrvdzpQpUygqKuKrr77CZrNx7bXXEhsby7XXXovL5eL222+nY8eOHD16lJtuuolp06YxdepUKioquOGGG5gwYQJjxoxhxYoV3Hrrrdxyyy0ArF69mtmzZ/Puu+9Sv3597r333mOeb/Hixbz11lu8+eablJeXM2TIEPr06cPLL79MdnY2I0eOpHHjxvTo0YPp06eTkZHBunXrKCkp4frrr692rYsuuohFixYRExPD0qVLuf322/nyyy9JTk7mySefZN68eSxevPi47/MXX3zhPiYlJYXHH3+cW2+9lUWLFrmPWblyJR9++CFHjx6lb9++XHHFFfTs2fOEP7u6osZ6Nv31r39l8uTJ3HPPPdx3330AHD16lMcff5w77riDxx9/nKNHj7qPf++997j99tu588472bp1q3v7nj17mDx5Mrfffjv/+c9/LpiqIIYJV/2hFEYPxqdkB2GZszE5irydSkRERERE5KzbsGEDY8eO5fnnnz9poSkzM5N169YxePBgoqOjufTSS4/p3dS+fXsuv/xyzGYzw4YNY9u2bQBs3ryZ4uJibrvtNnx8fLj00kvp1atXtV5U/fr1Iz09HYvFwpAhQ/jxxx8BSE9PJzg4mC+++AKoKkp17tyZ6Oho97njx48nOjqa+Ph4OnbsSHp6Oi1atMDX15d+/frxww8/ePReeHKvP/rrX/9KcHAwqamppKam0r17dy666CJCQkLo2bPnce+9ZcsWcnNzeeihhwgICMDPz48OHTq498fGxjJ+/HgsFgs+Pj4sXbqU+++/n6CgIOrVq8fEiRNZuHAhAA0bNqRbt274+voSGRnJhAkT+Prrr93ve2VlJTfccANWq5Urr7ySNm3auO+zdOlSrrnmGlJTU/H39z+mIAfQu3dv2rdvj8lkYtu2bRw5coRJkybh4+PDRRddxHXXXef+Of5WQAoLCyMhIYHx48dXu9bAgQOJi4vDZDIxaNAgGjZsWK0GcTLvvfceI0eOpGXLlvj6+rp7Qh04cMB9zG233UZoaCiJiYl07tzZ/Rmq62q0Z9MjjzxCSEiI+/XixYtp2bIlgwcPZvHixSxevJhRo0a5q47/+te/sNlsPP7447zwwguYTCZeffVVJk6c6K4ybt26lfT09Jp8DK8qC+2I0xJKSM48wjNmkJ8wlkqfWG/HEhERERGR89Tp9jiqCW+99RadOnWiS5cu7m2LFi1iypQpQNUwtblz57JgwQKSk5Np0aIFAEOGDGHq1Kk89NBDWK1WAGJiYtzX8Pf3p6ysDIfDQU5ODgkJCZhM/78PRlJSUrVhUL8v6Pj7+1NcXOx+PWLECBYuXEi3bt1YuHAhN954Y7VniIqKcv/dz8+v2rX8/PyqXetUTnWvP/rjvf/4+tChQ8eck5WVRVJSEhbL8csECQkJ7r/n5eVRUVFBUlKSe1tSUhLZ2dkAHD58mIceeogNGzZw9OhRnE4noaGhAOTm5hIXF4dhGMe9dm5uLq1atTruvuNty8jIIDc31z3MDap6aXXs2NF9vd8f/8frzZ8/n9mzZ5ORkQFAcXExeXl5x30P/ig3N7daL7rAwEDCw8PJycmhXr16wMk/Q3WZV+ds2rhxI927dwege/fubNy40b29c+fOWK1WYmJiiIuLY9euXdhsNkpLS0lJScEwDLp16+Y+50JSEdiU/MQJ4HIQnjELa+keb0cSERERERE5a5566ikyMzN55JFH3NuGDh3Kzp072blzJ3PnzgVgwYIF7N+/nzZt2tCmTRsee+wx8vLyWL169SnvERcXR1ZWlnueIajqKRUXF+dRxqFDh7J8+XJ+/PFHdu3aRZ8+fU7zKT1XE/dKSEggMzPzhJN//744FBERgdVqdRdooOq9i4+PB+DJJ5/EMAxWrFjBzz//zEsvveQelRQTE0NOTk61UUpZWVnuv8fExLiLVn/cd7wsCQkJ1KtXj23btrn/7Nixg7feeuuU18vIyODee+/ln//8Jz/88APbtm0jNTXVne339zme2NjYau9BSUkJNpvN489QXVajPZv++c9/AnDFFVfQq1cvCgoKCA8PByA8PJzCwkKgqkqanJzsPi8iIoK8vDzMZjORkZHu7ZGRkSesOK5cuZKVK1cCVb+ofl/JPV9ZLJbfPUcURD2I8fOLhGXNwdV4LER19GY8kVqjelsRkRNRWxHxjNqK1EWlpaXejnBSgYGBvP3221xzzTU88cQTPPDAA8cc88033/DLL7/wySefVPueOHXqVObPn0/v3r1Peo/09HQCAgKYMWMGEydOZOPGjaxYsYIPP/zQo4wJCQm0bt2aO++8kyuvvBJ/f//Te8gTiI6O5pdffqmRe/1eeno6MTExPPHEE/ztb3/DZDLx/fff0759+2OONZvNDBw4kKeffpoXXniB/Px8Zs+ezc033wxUTZkTEhJCaGgo2dnZzJw5031u27ZtMZlMzJkzh+uvv55PP/2UrVu3cskllwBVw9omT57M8OHDSUpKYvr06afMHRwczMsvv8z48ePx8fFh586dlJWV0aZNGwYOHMhLL71E69atKS0tZc6cOe5zS0pKMAyDiIgIoGoi9J9//tm9Pzo6muzsbCoqKvDx8Tnm3oMHD+bWW29l8ODBJCcn89RTT5Genu7u1VSXmEwm/P39Pf73sMaKTY8//jgREREUFBTwj3/847hd4X5zonmYTmd+pl69etGrVy/368OHD3setpaKioo65jmMuJsIzX4Ln12vcdR2gJKw7nCK6qtIXXe8tiIix1JbEfGM2orURbVp6foTCQ0NZd68eYwYMQKLxXLMRNHz58+nT58+1YZPAdxwww0MHToUm8120uv7+PgwZ84cHnjgAf79738TFxfHCy+84J4c3BMjRozgjjvuOGa1uDNxww03cNddd/HWW28xbNgwHn/88XN2r98zm828+eabPPTQQ7Rv3x7DMBgyZMhxi01Q9R3/oYce4pJLLsHX15e//OUvjBw5EoC7776bO++8k6ZNm9KgQQOGDRvmXj3Px8eH1157jb/97W88+eST9OzZk169ermLOZdddhnjx49nxIgRmEwm7rzzThYuXHjcYs9vud944w2mTp3KJZdcQkVFBY0aNXJ/XiZNmsSUKVO45JJLiImJYciQIfzvf/8DcK+GN2jQIAzDYPjw4dWe97cV/dLT0zEM45i5rrp27co999zDhAkTKCgooG3btsyYMeMMfgq1l9PppLS09Jh/D09U2zFcXphh+3//+x9+fn58+umnPProo4SHh2Oz2Xj00Ud54YUXeO+994Cq8bZQ1SNqxIgRREdH89hjj/H8888DVTO///TTT0yYMOGU9zxe17vzzQn/Q8flICR3AX5Hv6U0pANF0VeBYa75gCK1hL4UiHhGbUXEM2orUhcVFRURHBzs7Rjnva+//prbb7/dvXpbXblXTRswYACjR4/mmmuuOWbfzp07ueyyy9i7d+8J55M6HW+++Sbvv/++ezJz8czevXvZvHkzw4YNq7b9RMWmGvmElpWVubtplpWV8d1331G/fn3atWvHZ599BsBnn33mriC2a9eOdevWYbfbOXjwINnZ2TRp0oTw8HD8/f3ZsWMHLpeLtWvX0q5du5p4hNrNsFAYezXF4T3wL9xAaPZbGM5yb6cSERERERGps+x2O6+//jrXXXfdOS/+1OS9asJXX33FwYMHcTgc/O9//2Pbtm306NHDvf+jjz6ioqKC/Px8/vnPf3LFFVf86UJTbm4uGzduxOl0smvXLmbPnk3fvn3P0pPIidTIMLqCggKee+45oGpW+EsvvZQ2bdrQuHFjpk+fzqpVq4iKinIvaVivXj0uueQS7r77bkwmEzfccIO7Qd14443MmDGDiooK2rRpc0GtRHdShoniyD5UWsIIPrSEsMzZFMSPwWkJOfW5IiIiIiIi4rGdO3fSr18/mjVrxr/+9a86c6+asnv3bm6++WaKi4tp0KABr7zyCrGx/3+V9blz5zJp0iTMZjOdOnXiiSee+NP3stvtTJkyhf379xMaGspVV13FmDFjzsZjyEl4ZRidN9TpYXR/4FO8nZCcebjMAeQnjKXSJ/aU54jUJRruIOIZtRURz6itSF2kYXQicjpq5TA6qVkVgU3JT5wArkrCM2ZhLdnj7UgiIiIiIiIicoFQsamOcvglYku6Bac5hLCs/+BbtNXbkURERERERETkAqBiUx3mtIZjS7oZu/9FhOa+S0DeGrgwRk2KiIiIiIiIiJeo2FTHucz+5CeMoyyoNUF5nxB86D1wVXo7loiIiIiIiIjUUTWyGp14mWGhMPZqKq3hBNrWYHIUUBh3HS6Tr7eTiYiIiIiIiEgdo55NFwrDRHFkHwqjh+BTsouwzNmYHIXeTiUiIiIiIlJNx44dWbt2rbdj1JjMzEySk5OprNQIFKk7VGy6wJSFdqAgfjTmisOEZ8zEXJ7r7UgiIiIiIiJn5K677qJ+/frk5OR4O8ppS0xMZOfOnZjNZgCGDx/OO++8c0bXa926NQ6Hw73N4XDQqlUrEhMT3duGDx9Oo0aNSElJITU1lb59+/Lvf/+b8vLyY675x/c3Ozub+vXrs2/fvmOOveGGG5g6deqfzi91g4pNF6CKwKbkJ04AVyXhmbOwluz2diQREREREZE/paSkhA8//JDg4GDee+89b8epFUJCQli1apX79apVqwgNDT3muH/84x/s2LGDLVu28PDDD7NkyRJGjx6N63cLSx3v/Y2Pj+fSSy9l4cKF1a5ns9lYtWoVI0aMOEdPJucLFZsuUA6/RGxJt+C0hBCWNQffoi3ejiQiIiIiIlLNrl276NSpE0uWLDnhMcuWLSMkJIRJkyYxf/78avumTZvGxIkTueOOO0hJSaFnz558++237v07d+5k+PDhpKWl0bNnT5YvX+7ed9ddd/HAAw8wevRoUlJSGDBggLsnzwMPPMBjjz1W7V5jxozh1VdfBaqGAs6cOZNevXrRpEkTJk+ezKFDhxg1ahQpKSlcc8015OfnA3DgwAESExNxOBw89dRTrF+/nr///e8kJyfz4IMPnvJexzNs2DAWLFjgfj1//nyGDx9+wuMDAgLo3Lkzb7zxBps2bWLlypWnfH9HjBhxTLFpyZIlpKSkkJaWdsJ7yYVBE4RfwJzWcGyJNxOaM5fQ3P9x1J5PSXgPMAxvRxMRERERkRry2je57LWVndN7NAz348Z2sad1zvfff8/48eN54oknuOKKK0543Pz58xk8eDCDBg1i6tSpfP/997Rs2dK9f8WKFbz66qtMnz6dZ555hgcffJAPPvgAu93O2LFjueaaa3jnnXfYsGED48eP58MPP6RJkyZAVfFk7ty5tGzZkrvuuounn36amTNnMmLECG644QYeeughTCYTeXl5fPHFFzz33HPu+y5btox58+bhcDjo06cPP/zwA9OmTSM5OZlRo0bxn//8h7vvvrvas9x333188803DB06lOuuuw6ALVu2nPJef9S3b1/mzJlDQUEBABs2bGDy5Mk888wzJ33PfxuCt2HDBvd7fqL3t2/fvtx///1s2LCBDh06ALBw4cKTFrXkwqGeTRc4l9mf/IRxlAW1IShvOcGH3gOXJqYTERERERHv2bBhA2PHjuX5558/aaEpMzOTdevWMXjwYKKjo7n00kuP6d3Uvn17Lr/8csxmM8OGDWPbtm0AbN68meLiYm677TZ8fHy49NJL6dWrV7VeVP369SM9PR2LxcKQIUP48ccfAUhPTyc4OJgvvvgCqCpKde7cmejoaPe548ePJzo6mvj4eDp27Eh6ejotWrTA19eXfv368cMPP3j0Xnhyrz/y9fXliiuu4P3332fJkiX07t0bX1/PViOPjY1197o62fvr7+/PgAED3D2o9uzZw/fff8+QIUM8uo/UberZJGBYKIy9mkprOIG21ZgcBRTGXYfL5NkvIxEREREROX+dbo+jmvDWW2/RqVMnunTp4t62aNEipkyZAlQNU5s7dy4LFiwgOTmZFi1aADBkyBCmTp3KQw89hNVqBSAmJsZ9DX9/f8rKynA4HOTk5JCQkIDJ9P/7YCQlJVWbZPz3BR1/f3+Ki4vdr38bRtatWzcWLlzIjTfeWO0ZoqKi3H/38/Ordi0/P79q1zqVU93reIYPH86TTz6Jy+XiwQcf9PheOTk5tGvXDuCU7++IESMYO3YsU6dOZeHChfTo0aPac8uFSz2bpIphUBzZm8LoIfiU7CIsczYmR6G3U4mIiIiIyAXoqaeeIjMzk0ceecS9bejQoezcuZOdO3cyd+5coKoYsn//ftq0aUObNm147LHHyMvLY/Xq1ae8R1xcHFlZWTidTve2zMxM4uLiPMo4dOhQli9fzo8//siuXbvo06fPaT6l5/7MvTp27MjBgwc5fPiwe5jbqWRmZvLdd9+5jz/V+9uxY0fCw8P55JNPWLRokYbQiZuKTVJNWWgHCuKvx1xxmPCMGZjLc70dSURERERELjCBgYG8/fbbrF+/nieeeOK4x3zzzTf88ssvfPDBByxfvpzly5ezatUqhgwZcsxQuuNJT08nICCAGTNmYLfbWbduHStWrGDQoEEeZUxISKB169bceeedXHnllfj7+5/WM55IdHQ0v/zyyxnfyzAM3njjDebMmYNxinl5S0tL+eqrrxg/fjzp6elcfvnlHr+/w4cP54knnqCwsPCkQx7lwqJikxyjIjCV/MQJ4HISnjkLa8lub0cSEREREZELTGhoKPPmzWPVqlXHndh6/vz59OnTh7S0NGJiYtx/brjhBlauXInNZjvp9X18fJgzZw6rV6+mZcuWPPDAA7zwwgvuycE9MWLECLZt28awYcNO+/lO5IYbbmDZsmU0a9aMhx566IzulZqaSmpq6gn3//3vfyclJYXWrVvzyCOPcOWVVzJ37lxMJpPH7+/w4cPJzMxk4MCBHs8LJXWf4XK5XN4OUROysrK8HeGMRUVFcfjw4Rq7n8luIyz7DcwVRyiMHUZ5cHqN3VvkTNR0WxE5X6mtiHhGbUXqoqKiIoKDg70d47z39ddfc/vtt7N+/fpqcz+d7/cS+aO9e/eyefPmY4qdCQkJxz1en1A5Iac1HFvizdj9LyI0938E5K2GC6M2KSIiIiIiclJ2u53XX3+d66677pwXf2ryXiJngz6lclIusz/5CeMoC2pDUN5ygg+9B65Kb8cSERERERHxmp07d5KWlkZubq5HK8OdL/cSOVss3g4g5wHDQmHs1VRawwm0rcbkKKAw7jpcJo3HFRERERGRC09ycjK7du2qc/cSOVvUs0k8YxgUR/amMHooPiW7CMt4BZOj0NupRERERERERKSWUbFJTktZaHsK4q/HbD9CeMYMzOW53o4kIiIiIiIiIrWIik1y2ioCU8lPmgguJ+GZs7CW7PZ2JBERERERERGpJVRskj/F4ZuALelWnJZQwrLm4Fu0xduRRERERERERKQWULFJ/jSnNQxb4kTs/hcRmvs/AvJWgcvl7VgiIiIiIiIi4kUqNskZcZn9yU8YR1lwG4LyVhB86D1wVXo7loiIiIiInKc6duzI2rVra+x+iYmJ7N27t8buJ3IhULFJzpxhoTDmaorDe+JfuJHQ7DcxnOXeTiUiIiIiInXYtGnTuP322095XG5uLpMnTyY9PZ2UlBS6devGc889R0lJSQ2kFLkwqdgkZ4dhUBzZm8LoofiU7CYs4xVMjkJvpxIRERERkQuYzWbjqquuoqysjPfff58dO3Ywb948CgsL2bdvn7fjidRZKjbJWVUW2p6C+Osx248QnjEDc3mOtyOJiIiIiMh5ateuXXTq1IklS5ZU27569Wpeeukl3n//fZKTk+nVq9dxz589ezZBQUG89NJL1KtXD6gaNjd16lSaNWt2zPGFhYXccccdtGzZkg4dOvD888/jdDoB2Lt3L8OGDaNp06a0aNGCm2++uVrOkSNH0rx5c7p27cr777/v3nfXXXfxwAMPMHr0aFJSUhgwYEC1QtfJzv3000/p0aMHKSkptG3bllmzZgGQl5fH9ddfT1paGs2bN2fIkCHunCK1gcXbAaTuqQhMJT9pIqFZbxCeOYuCuNHYAxp7O5aIiIiIiBzHD5tLKMw/t/OuhoSZaXFxwGmd8/333zN+/HieeOIJrrjiimr7evbsye23386+fft46aWXTniNzz//nH79+mEyedbP4u9//ztFRUV89dVX2Gw2rr32WmJjY7n22mt59tln6datG/Pnz6eiooLvvvsOgJKSEkaOHMk999zD3Llz2bZtG9deey2pqamkpqYCsGTJEubOnUvLli256667ePrpp5k5c+Ypz/3b3/7GrFmz6NixI/n5+Rw4cACAV155hfj4eHeGzZs3YxjGab2/IueSejbJOeHwTcCWdCtOSxhhWXPwK9zs7UgiIiIiInKe2LBhA2PHjuX5558/ptB0Omw2GzExMR4dW1lZydKlS7n//vsJCgqiXr16TJw4kYULFwJgsVjIzMwkJycHPz8/OnToAMCKFSuoV68e11xzDRaLhZYtW3LllVeybNky97X79etHeno6FouFIUOG8OOPP3p0rsViYceOHRQVFREWFkbLli3d2w8ePEhGRgZWq5WOHTuq2CS1ino2yTnjtIZhS5xIaM5cQg7Ox+TIpyS8J+iXoIiIiIhIrXG6PY5qwltvvUWnTp3o0qULAIsWLWLKlClA1Wp1c+fOPeac9evXM2rUKACSkpJYvXo14eHhHDx40KN75uXlUVFRQVJSkntbUlIS2dnZQFWvp2effZYBAwYQGhrKxIkTGTlyJJmZmWzZsoW0tDT3eQ6Hg2HDhrlfR0dHu//u7+9PcXExwCnPffXVV3nhhRd48sknSUtL4/7776ddu3bccsstTJs2jeuuuw6Av/zlL9x2220ePadITVCxSc4pl9mf/IRxBB9cRFDeCswOG0XRg8EwezuaiIiIiIjUUk899RQvv/wyjzzyCI899hhDhw5l6NCh1Y75Y0+ejh07snPnzmrbunbtyscff8zdd999yqF0ERERWK1WMjIySElJAaqKQfHx8QDExMTw7LPPAlU9r0aOHEnHjh1JSEigU6dO/N///d9pP+epzm3Tpg1z5szBbrczZ84cbr75Zr755huCgoJ45JFHeOSRR/j5558ZMWIErVu3pmvXrqedQeRc0DA6OfcMC0UxIygOvwz/wm8IzX4Tw1nm7VQiIiIiIlJLBQYG8vbbb7N+/XqeeOKJ4x4TFRXFgQMHTjox9oQJEygqKuLOO+8kIyMDgOzsbB599FF++umnaseazWYGDhzI008/zdGjR8nIyGD27NnuItfSpUvJysoCIDQ0FMMwMJvN9OrViz179rBgwQLsdjt2u52tW7ceU/g6npOdW1FRwaJFiygsLMRqtRIcHIzZXPU/7VesWMHevXtxuVwEBQVhNpvd+0RqAxWbpGYYBsWRV1AYMxSfkt2EZczG5Cj0dioREREREamlQkNDmTdvHqtWreKZZ545Zv+AAQMAaNGiBX369DnuNcLDw1myZAlWq5UBAwaQkpLCNddcQ0hICA0bNjzm+Mcff5yAgAAuueQSBg8ezJAhQxg5ciQA3377LQMHDiQ5OZlx48bx2GOPUb9+fYKCgnjnnXdYsmQJF198MW3atOGf//wn5eXlp3zGU527cOFCOnXqRGpqKm+99ZZ7MvS9e/cycuRIkpOTueqqq7j++uvp3LmzZ2+sSA0wXC6Xy9shasJvFejzWVRUFIcPH/Z2jDPmU7yDkJy3q4bYxY+l0jfO25GkjqkrbUXkXFNbEfGM2orURUVFRQQHB3s7hoicJ/bu3cvmzZurzUUGVUNBj0c9m6TGVQSmkJ80EVwuwjNnYS3Z5e1IIiIiIiIiInKWqNgkXuHwTcCWdAtOSxhhWXPwK9zs7UgiIiIiIiIichao2CRe47SGYUuciN2/ISEH5xOQ9ylcGKM6RUREREREROosFZvEq1xmf/ITxlIanE5Q3kqCDy4CV6W3Y4mIiIiIiIjIn2TxdgARDAtFMSNwWsIJtK3C5CigMP46XCY/bycTERERERERkdOknk1SOxgGxZFXUBgzDJ/S3YRlzMbkKPB2KhERERERERE5TSo2Sa1SFtKOgvgxmO1HCM+Yibk8x9uRREREREREROQ0qNgktU5FYAr5SRPB5SI8cxbWkl3ejiQiIiIiIiIiHlKxSWolh28CtqRbcFrCCMuag1/hJm9HEhERERGRWmLKlClMnz7d/frNN9+kdevWJCcnk5eXx8aNG+nSpQvJycl8/PHHXkxapWfPnqxbt86jYxMTE9m7d+9ZuW9mZibJyclUVp6fizCdzvsmsH79erp27ertGAAYLteFsdZ8VlaWtyOcsaioKA4fPuztGDXKqCwjNGcuPqW7ORrRi5Lwy8AwvB1LarkLsa2I/BlqKyKeUVuRuqioqIjg4GBvxziujh07cvjwYcxmM2azmeTkZIYPH86oUaMwmY7tL2G322natCnvv/8+zZs3B+Dqq6+md+/e3HjjjTUdn7vuuov4+HimTJnyp85PTEzkiy++oGHDhmc5mdQ1NflZ2bt3L5s3b2bYsGHVtickJBz3ePVsklrNZfYjP2EspcEXE5S3kuCDC8F1flblRURERETEM3PmzGHHjh2sX7+e2267jRkzZjB58uTjHnvo0CHKyspITU11b8vMzCQlJeVP3dvhcPyp82qbuvIcp6s2PXdtylLTVGyS2s+wUBQznOLwy/Av2kRY1hsYzjJvpxIRERERkXMsJCSE3r17M3PmTObPn8/27duBqt5DTz/9NLt376Zbt24ApKWlMWLECDp37swvv/zCuHHjSE5Opry8nMLCQiZPnkx6ejpt27bl6aefdg8te/fddxk0aBCPPPIIzZs3Z9q0aZSXlzN16lTat29P69atmTJlCqWlpQCsW7eOtm3bMmvWLFq1akV6ejrvvvsuAHPnzuW9995j5syZJCcnM2bMGKCqt9batWsB2LJlCwMHDiQtLY309HQefPBBKioqPHo/hg8fzpNPPkn//v1p2rQp48aNw2azAXDgwAESExOZN28e7du35+qrr3ZvczgcLFmyhH79+lW73uzZsxk7diwAK1eupHfv3qSmptKuXTumTZtW7dgNGzZw1VVXkZaWRrt27Xj33XfZunUrrVu3rlZUWbZsGVdcccVx8991113cf//9jBo1iuTkZAYNGsTBgwd5+OGHadasGd26deOHH35wH//7962yspIXX3yRzp07k5KSQt++fcnMzASqevi88cYbdOnShUsvvRSAt99+my5dutC8eXPGjh1LTs7/X3zq4Ycfpl27dqSmptK3b1/Wr1/v3ldaWsqdd95Js2bN6N69OzNmzKBt27bu/d9//z29e/cmJSWFCRMmcPPNN/P0008D//+z8fLLL9OmTRsmTZqE0+nk3//+N507d6Z58+ZMnDjR/TMDmD9/Ph06dKB58+ZMnz7d48/K0KFDAbjiiitITk5myZIl7vv/ZufOnQwfPpy0tDR69uzJ8uXLq/0sHnjgAUaPHk1KSgoDBgxg3759x/25/RmWs3YlkXPJMCiOvIJKazjBB98jLOMVChLG4rSEejuZiIiIiMh5be3atRw6dOic3iM6OtpdFPoz0tPTiY+PZ/369TRt2tS9vXHjxqxevZpOnTqxbds2LJaqr7gdO3bk2Wefdd/zlltuISoqii+//JKSkhLGjBlDQkICo0ePBqq+1A8aNIhvv/0Wu93OP//5T/bv38/y5cuxWq389a9/5fnnn+f+++8HqnpTFRUVsWnTJtauXcuECRPo06cPo0aN4ptvvjnpMDqz2cyjjz5K69atyc7OZtSoUbz55pvcdNNNHr0XCxYs4O2336Z+/frceeedPPzww7z00kvu/V999RWfffYZhmFUGwLcu3dv7rnnHvbs2UOjRo0AWLx4MRMnTgQgICCAF154gdTUVLZv3861115L8+bN3UWdUaNG8cwzz9C/f3+KiorIysqiRYsWhIeHs3btWi677DIAFi1adMxQq99bunQp77zzDikpKYwePZqrrrqKyZMn88gjj/Dcc8/x6KOPsmDBgmPOmz17NkuWLOG///0vjRs35qeffsLf39+9/+OPP+aDDz7Az8+PL774gieffJJ58+aRkpLC448/zq233sqiRYsAaN26NXfddRchISG89tprTJw4ka+//ho/Pz+mT59ORkYG69ato6SkhOuvv959j4qKCm644QYmTJjAmDFjWLFiBbfeeiu33HKL+5hDhw6Rn5/P+vXrcTqdvP7663z88ccsWLCAyMhIHnroIR588EFmzJjBjh07eOCBB3j77bdp06YNTz31VLWi2Mk+K4sWLSIxMZEVK1a4h9H9fn4ru93O2LFjueaaa3jnnXfYsGED48eP58MPP6RJkyYALFmyhLlz59KyZUt3AXfmzJkn/NmdDvVskvNKWUg7ChLGYLbbCM+Yibk859QniYiIiIjIeS82Npb8/PzTPu/QoUOsXr2axx57jICAAKKiorjppptYsmRJtWuPHz8ei8WCn58f77zzDo8++ijh4eEEBQVx++23VzvearUyadIkrFYrl19+OYGBgezevdujPK1ataJt27ZYLBbq1avHqFGj+Prrrz1+nmHDhtG0aVMCAgK45557WLp0abUJwCdPnkxAQEC1QgyAv78/ffr0cT/Hnj172LVrF7179wagc+fOpKWlYTKZaNasGYMGDeKrr74CqgpIXbt2ZfDgwVitViIiImjRogUAI0aMcBdxbDYba9asYciQISfM369fP1q1aoWfnx/9+vXD19eXESNGYDabueqqq/jxxx+Pe94777zDvffeS5MmTTAMg+bNmxMREeHef9tttxEeHo6/vz/vvfceI0eOpGXLlvj6+nL//fezadMmDhw44H4PIyIisFgs3HzzzVRUVLh/fkuXLuX2228nLCyMhIQExo8f777H5s2bqays5IYbbsBqtXLllVfSpk2bajlNJhOTJ0/G19cXf39/5s6dy5QpU0hISMDX15fJkyezbNkyHA6HuxdYhw4d8PHx4Z577sH43RzFZ/JZ2bx5M8XFxdx22234+Phw6aWX0qtXr2qf4379+pGeno7FYmHIkCEnfO//jBrt2eR0OrnvvvuIiIjgvvvu4+jRo0yfPp1Dhw4RHR3NpEmTCAoKAuC9995j1apVmEwmxo0b5/4B7tmzh5dffpmKigrS09MZN25ctR+G1H0VASnkJ00kNOsNwjNnURA3CntAE2/HEhERERE5L51Jj6OalJOTQ1hY2Gmfl5GRgd1u5+KLL3Zvczqd1SY2/v3fjxw5QmlpabUhZy6Xq1pBJywszN2LCqoKOcXFxR7l2b17N4899hjfffcdpaWlOBwOWrVq5fHz/D5rUlISdrudvLy84+7/o8GDBzN16lQmTZrE4sWL6dOnj7sotXnzZp544gl+/vln7HY7FRUV9O/fH6hacOuiiy467jWHDh1Kjx49KC4uZunSpXTs2JHY2NgTZoiKinL/3c/Pj+jo6GqvT/Q+nizDH587NzeXli1bul8HBgYSHh5OTk4O9erVY9asWcybN4/c3FwMw6CoqMj9Hubm5p7ws5Gbm0tcXFy1GsQf3+/IyEj8/PzcrzMyMrjxxhurTW5vNps5dOgQOTk51c739/cnPDzc/fpMPiu/Xfv3901KSqrWc+r37/3pfIY9UaM9mz788EMSExPdrxcvXkzLli158cUXadmyJYsXLwZwd1n717/+xYMPPsjrr7+O0+kE4NVXX2XixIm8+OKL5OTksHXr1pp8BKklHL7x2JJuwWkJJyxrDn6Fm7wdSUREREREzpGtW7eSk5NDhw4dTvvc33qUfP/992zbto1t27bx888/s3r1avcxvy8eRERE4Ofnx6pVq9zHb9++nZ07d3p0v1N1hrj//vtp0qQJX3zxBT///DP33Xcfp7NI/O9XWs/MzHT3NPLk/t27dycvL48ffviBxYsXV+uBdNttt9G7d282btzI9u3bGTVqlHtfQkICv/zyy3GvGR8fT9u2bfnoo49YuHDhSYfQnYmTZYDqzx0bG0tGRob7dUlJCTabjbi4ONavX8+MGTOYNWsWP/30E9u2bSMkJMT9M4iJiSE7O9t97u/f75iYGHJycqr9vH6//0S5586d6/4sbdu2jT179hAfH09sbGy1e5WWllabz+lMPitxcXFkZWW5aylQ9XmJi4vz6PwzVWPFpiNHjrB582Yuv/xy97aNGzfSvXt3oOpDv3HjRvf2zp07Y7VaiYmJIS4ujl27dmGz2SgtLSUlJQXDMOjWrZv7HLnwOK1h2BInYvdvRMjBBQTkfQqn8UtaRERERERqt6KiIve8OEOHDiUtLe20rxEbG0u3bt2YOnUqRUVFOJ1O9u3b5x4i9kcmk4m//OUvPProo+45j7Kzs1mzZo1H94uOjmb//v0n3F9cXExwcDCBgYHs2rWL//73v6f1PAsXLmTHjh2Ulpby7LPP0r9/f8xms0fnWiwW+vfvzz/+8Q/y8/Or9Wo7evQoYWFh+Pn5sWXLFndnEKjqvfT555/z/vvv43A43AWr3wwfPpyZM2eyffv2YyYhP1uuu+46nn32Wfbs2YPL5eKnn36q1qPr9wYPHsy7777LDz/8QHl5OU899RTp6enUq1ePo0ePYrFYiIyMxOFwMH36dIqKitznDhw4kJdeeon8/Hyys7OZM2eOe1/btm0xmUzMmTMHh8PBJ598csoOMKNHj+bpp592F7+OHDnCJ598AkD//v1ZsWIFGzdupKKigmnTplUrJp3qs3Kyz1p6ejoBAQHMmDEDu93OunXrWLFiBYMGDTpp3rOlxopNb7zxBqNGjapWbSwoKHB3EQsPD6ewsBCAvLw8IiMj3cdFRESQl5d3zPbIyMgTfrjkwuAy+5GfMIbS4IsJyltJ8MGF4Ko89YkiIiIiIlJrjRs3jpSUFNq3b8+LL77IhAkTmD59+p++3gsvvEBFRQU9evSgWbNmTJgwgYMHD57w+AceeIAGDRowcOBAUlNTGTlypMdzMo0cOZIdO3aQlpZWbb6f3zz00EO89957pKSkcM8993DVVVed1rMMHz6cSZMm0aZNG8rLy3n88cdP6/whQ4bw+eefM2DAgGpDAZ944gmee+45UlJSmD59OgMHDnTvS0xM5K233mL27Nk0b96c3r1789NPP7n39+vXj4yMDPr27UtAQMBp5fHUhAkTGDBgANdddx2pqan87W9/o6zs+KuUd+3alXvuuYcJEyZw8cUXs2/fPmbMmAFAjx496NmzJ127dqVjx474+vpWG8o2adIk4uPjueSSSxg5ciT9+/fH19cXAB8fH1577TXmzZtHWloaCxcupFevXvj4+Jww94033sgVV1zBtddeS0pKCgMHDmTz5s0ApKamuicvT09PJzAwkKioKPf9TvVZufvuu7nrrrtIS0vj/fffr7bPx8eHOXPmsHr1alq2bMkDDzzACy+84J4c/FwzXKfTX+9P2rRpE1u2bOHGG2/kxx9/ZOnSpdx3332MHTuWN954w33cuHHjmDNnDq+99hopKSnuKuvMmTNJT08nKiqKefPm8dBDDwGwbds2lixZwn333XfMPVeuXMnKlSsBeOqppzxeSrI2s1gs1ZaUlN9xuTAylmJkLsUV2gxX8s1g8T/1eVInqa2IeEZtRcQzaitSFx04cOCYCaSl9hs+fDhDhw7luuuu83aUY3Tu3JmnnnrqvJkDzFNvvvkm77//PgsXLjzu/gEDBjB69GiuueaaM75XcXExaWlpfPHFF9SvX/+Mr3c2/fLLL/z444/VVucDTlhoq5EJwn/++We++eYbtmzZQkVFBaWlpbz44ouEhoZis9kIDw/HZrMREhICVPVYOnLkiPv8vLw8IiIijtl+5MiRamNTf69Xr1706tXL/fr3Sz6er6KiourEc5wz/p3xi/Eh+OB7VH73BAUJY3FaQr2dSrxAbUXEM2orIp5RW5G6SAVUOZuWLVuGYRhceuml3o5yxnJzc9m/fz9t27Zlz549zJ49m7Fjx7r3f/XVVzRu3JiIiAgWLVrEtm3b6NGjx5++3/Lly+natSsul4upU6fStGlT6tWrd+YPcpY5nU5KS0uP+ffwRBPS10ix6brrrnNXXn/r2XTHHXfw1ltv8dlnnzF48GA+++wz2rdvD0C7du148cUXGTBgADabjezsbJo0aYLJZMLf358dO3aQnJzM2rVr6du3b008gpwnykLaUWkJJTT7bcIPzCA/YSyVvvHejiUiIiIiIlInDR8+nB07dvDiiy9WW/nsfGW325kyZQr79+8nNDSUq666ijFjxrj37969m5tvvpni4mIaNGjAK6+8ctLV905l+fLl3HnnnbhcLlq1asWMGTNOOcn8+aBGhtH93u+H0RUVFTF9+nQOHz5MVFQUd999N0FBQQAsWrSI1atXYzKZGDt2LOnp6UDVD3bGjBlUVFTQpk0bxo8f79EP4lQzxJ8P9H/VPGcuzyYs6w0MZzkF8X/BHpDs7UhSg9RWRDyjtiLiGbUVqYuKiooIDg72dgwROU/s3buXzZs3H7Pa4Il6NtV4sclbVGy68JgcBYRlvYG54iBFMUMpC2nr7UhSQ9RWRDyjtiLiGbUVqYtUbBKR03G6xabzv4+byAk4LaHYEidi929EyMEFBOSthAujtioiIiIiIiLiNSo2SZ3mMvuRnzCW0uC2BOV9SvDBheCq9HYsERERERERkTqrRiYIF/Eqw0xRzDAqrWEE5X2K2VFAQfxfcJn8vJ1MREREREREpM5Rzya5MBgGJRG9KIwZjrV0D2EZr2ByFHg7lYiIiIiIiEido2KTXFDKQtqSnzAOs91G+IEZmMuzvR1JRERERERO05QpU5g+fbr79Ztvvknr1q1JTk4mLy+PjRs30qVLF5KTk/n444+9mLRKz549WbdunUfHJiYmsnfv3rNy38zMTJKTk6ms1FQiUrNUbJILjj2gCbakiQCEZ7yCtWSnlxOJiIiIiMhvOnbsSOPGjUlJSSEtLY2rrrqK//73vzidTvcxTz/9NJMmTQLAbrczdepU3nnnHXbu3ElERATPPvss48aNY+fOnfTt27dG89911108/fTT1batXr2azp0712gOqCpc7dy5E7PZXOP3Pp677rqLxMREli9fXm37ww8/TGJiIu+++y4A7777LvXq1SM5OZnk5GQ6derEpEmT2L179zHXXLduHYmJicyYMcO97brrruPZZ5895thPPvmENm3a4HA4zvKTyR+p2CQXpErfeGz1bqXSGk5Y1hv4FX7j7UgiIiIiIvKrOXPmsGPHDtavX89tt93GjBkzmDx58nGPPXToEGVlZaSmprq3ZWZmkpKS8qfuXVcKEbX1ORo1asT8+fPdrx0OB8uWLaNBgwbVjmvbti07d+5k+/bt/N///R9+fn7069eP7du3Vztu/vz5hIWFVbvm1VdfzcKFC3H9YTXyBQsWMGTIECwWTV99rqnYJBcspyWU/KSJ2P0bEXJwIYFHVsIffhmJiIiIiIj3hISE0Lt3b2bOnMn8+fPdhYbfeg/t3r2bbt26AZCWlsaIESPo3Lkzv/zyC+PGjSM5OZny8nIKCwuZPHky6enptG3blqeffto9tOzdd99l0KBBPPLIIzRv3pxp06ZRXl7O1KlTad++Pa1bt2bKlCmUlpYCVT1p2rZty6xZs2jVqhXp6enuHjlz587lvffeY+bMmSQnJzNmzBigqrfW2rVrAdiyZQsDBw4kLS2N9PR0HnzwQSoqKjx6P4YPH86TTz5J//79adq0KePGjcNmswFw4MABEhMTmTdvHu3bt+fqq692b3M4HCxZsoR+/fpVu97s2bMZO3YsACtXrqR3796kpqbSrl07pk2bVu3YDRs2cNVVV5GWlka7du1499132bp1K61bt65W2Fq2bBlXXHHFCZ/hiiuuYOPGjeTn5wNVvb7S0tKIjo4+7vFms5kGDRrw5JNP0qlTp2q5SktLWbZsGU888QR79+7l22+/BaBPnz7k5+ezfv1697H5+fl8+umnDB8+/BTvspwNKufJBc1l8iM/YSzBB98j0PYpJoeNopghYKhpiIiIiMiFIejQUizneC5Th288R6MH/unz09PTiY+PZ/369TRt2tS9vXHjxqxevZpOnTqxbds2d4+Vjh078uyzz7oLUbfccgtRUVF8+eWXlJSUMGbMGBISEhg9ejRQVQAaNGgQ3377LXa7nX/+85/s37+f5cuXY7Va+etf/8rzzz/P/fffD1T1pioqKmLTpk2sXbuWCRMm0KdPH0aNGsU333xDfHw8U6ZMOe6zmM1mHn30UVq3bk12djajRo3izTff5KabbvLovViwYAFvv/029evX58477+Thhx/mpZdecu//6quv+OyzzzAMg8OHD7u39+7dm3vuuYc9e/bQqFEjABYvXszEiVVTjAQEBPDCCy+QmprK9u3bufbaa2nevDl9+/YlMzOTUaNG8cwzz9C/f3+KiorIysqiRYsWhIeHs3btWi677DIAFi1axLBhw06Y39fXl969e7NkyRLGjBnDggULGD58OG+88cYpn/3KK6/kySefdL9etmwZgYGBDBgwgEWLFrFgwQJat26Nv78/AwcOZMGCBXTq1AmApUuX0rhxY5o3b+7R+yxnRj2bRAwzRTHDOBrRC/+izYRlvYlRWebtVCIiIiIi8juxsbHu3jCn49ChQ6xevZrHHnuMgIAAoqKiuOmmm1iyZEm1a48fPx6LxYKfnx/vvPMOjz76KOHh4QQFBXH77bdXO95qtTJp0iSsViuXX345gYGBx51P6HhatWpF27ZtsVgs1KtXj1GjRvH11197/DzDhg2jadOmBAQEcM8997B06dJqE4BPnjyZgIAA/P39q53n7+9Pnz593M+xZ88edu3aRe/evQHo3LkzaWlpmEwmmjVrxqBBg/jqq6+AqgJS165dGTx4MFarlYiICFq0aAHAiBEjWLRoEQA2m401a9YwZMiQkz7D8OHDWbBgAYWFhXz99df06dPHo2f/42dg/vz5DBw4ELPZzJAhQ1i8eDF2u92d64MPPnD3SFuwYAEjRozw6D5y5tR9QwTAMCiJuBynJYzgg4sIz3yF/ISxOC2h3k4mIiIiInJOnUmPo5qUk5NDWFjYaZ+XkZGB3W7n4osvdm9zOp0kJCS4X//+70eOHKG0tLTakDOXy1WtoBMWFlZt3h9/f3+Ki4s9yrN7924ee+wxvvvuO0pLS3E4HLRq1crj5/l91qSkJOx2O3l5ecfd/0eDBw9m6tSpTJo0icWLF9OnTx93UWrz5s088cQT/Pzzz9jtdioqKujfvz8AWVlZXHTRRce95tChQ+nRowfFxcUsXbqUjh07Ehsbe9Jn6NChA3l5ebzwwgv06tXrmMLYifz+M5CZmcm6devcvc369OnDvffey6effkrfvn3p0KEDkZGRLF++nDZt2vDtt9/y2muveXQfOXMqNon8TllIWyotoYRmzyX8wAwKEsbi8I33diwRERERkQva1q1bycnJoUOHDqd9bkJCAr6+vnz//fcnnBjaMAz33yMiIvDz82PVqlXEx5/+d4HfX+t47r//flq0aMGMGTMICgri1VdfZdmyZR5fPysry/33zMxMd0+j37af7P7du3cnLy+PH374gcWLF/Poo4+69912222MHTuWuXPn4ufnx8MPP+yeDyohIYGtW7ce95rx8fG0bduWjz76iIULF3L99dd79BxDhw5l+vTp1Sb2PpWPPvrI/RlYuHAhTqfTPecUQHl5OQsWLHCvQPhbD6rf5vY60bxQcvZpGJ3IH9gDmmBLqhq3HJbxCj4lO7ycSERERETkwlRUVMSKFSu49dZbGTp0KGlpaad9jdjYWLp168bUqVMpKirC6XSyb98+9xCxPzKZTPzlL3/h0Ucfdc95lJ2dzZo1azy6X3R0NPv37z/h/uLiYoKDgwkMDGTXrl3897//Pa3nWbhwITt27KC0tJRnn32W/v37YzabPTrXYrHQv39//vGPf5Cfn++e0wrg6NGjhIWF4efnx5YtW1i8eLF739ChQ/n88895//33cTgc7oLVb4YPH87MmTPZvn37MZOQn8j48eOZN2+ee06lE6msrGT//v38/e9/56uvvuLuu+8GqobF3X333Sxfvtz959VXX+XTTz919/QaPnw4n3/+OW+//baG0NUwFZtEjqPSNx5bvVuptIYTmvUmfoXfeDuSiIiIiMgFY9y4caSkpNC+fXtefPFFJkyYwPTp0//09V544QUqKiro0aMHzZo1Y8KECRw8ePCExz/wwAM0aNCAgQMHkpqaysiRIz2ek2nkyJHs2LGDtLQ0xo8ff8z+hx56iPfee4+UlBTuuecerrrqqtN6luHDhzNp0iTatGlDeXk5jz/++GmdP2TIED7//HMGDBhQrafXE088wXPPPUdKSgrTp09n4MD/P7wyMTGRt956i9mzZ9O8eXN69+7NTz/95N7fr18/MjIy6Nu3LwEBAR7lCA8Pp2vXrifsibVp0yaSk5NJTU1l+PDhFBUV8eGHH5KWlsamTZvIyMhg7NixxMTEuP/07t2bBg0auOelqlevHu3ataO0tNQ9N5XUDMPlujDWev99V8PzVVRUVLXVBOTcM5xlhGS/g2/pTorDL6M4ohecoluseJ/aiohn1FZEPKO2InVRUVERwcHB3o4hp2n48OEMHTqU6667zttRjtG5c2eeeuqpar2lpO7Yu3cvmzdvPmalwRPNEaaeTSIn4TL5UZAwhtLgtgTaVhF8cAG4HN6OJSIiIiIiUmssW7YMwzC49NJLvR1FaglNEC5yKoaZophhVFrDCcpbidlRQEHcKFxmP28nExERERER8arhw4ezY8cOXnzxRUwm9WeRKio2iXjCMCiJuBynJYzgg4sIz5xFfvxYnNYwbycTEREREZELxIIFC7wd4Ri1MZN4n8qO5wmXy8UvthJvx7jglYW0JT9hHCZHPuEZM7GUZ3s7koiIiIiIiEitomLTeeLzX4oY9dZmXtuUS4m90ttxLmj2gCbYEm8GwyAs4xV8SnZ4O5KIiIiIiIjIOXO6a8up2HSeuDg+kKtaxPHBdhu3Ld3Luv2Fp/3DlrOn0jcOW9ItVFojCM16E7/Cb7wdSURERETEY4ZhUF5e7u0YIlLLuVwubDYbeXl5p3We5mw6TwT5mrnnsiZ0TvBl5oYcnv48i7YJgUxsH0tskI+3412QnJZQ8pMmEJL9DiEHF2K22yiO6AWG4e1oIiIiIiInFRgYyK5duzCZTGdtUmeTyYTT6Twr1xKpy86XtuJyuaisrOTAgQMUFBRgsXheQlKx6TyTGuXPtL4N+OBnG+98d4jbPtjLNS2iGJQWgdWsIkdNc5n8KEgYQ/DBxQTaVmFy2CiKGQqGmpaIiIiI1F6GYZCYmMgnn3yCw+HAbDaf8TUDAgIoKdE8syKncr61FafTicPhoFOnTh6fo2/E5yGzyWBQWgSd6wfz+qZc3vr2EGv2FXBL+ziaxwZ4O96FxzBTFDOUSmsYQXkrMTsKKYgbhcvs5+1kIiIiIiInFBAQQJ8+fdi9ezelpaVnPE1HSEgIhYWFZymdSN11vrUVi8VCvXr1iImJ8fgcw3WBTPyTlZXl7QhnLCoqisOHDx+zfWPGUWZ/k8PBYgeXNwplbHo0IX6qI3qDX+Fmgg8upNInmvz4sTitYd6OdEE6UVsRkerUVkQ8o7Yi4hm1FRHP1KW2kpCQcNztmiC8DmifFMRLAxoxtFkEa/YWcOvSPazcnY/zwqgj1iplIReTnzAOkyOf8IyZWMrP/yKniIiIiIiIyOlQsamO8LOYGJMew/QrG1Iv1JeXvs7hwRX72Z+vFSZqmj2gCbbEm8EwCMt4BZ/iHd6OJCIiIiIiIlJjVGyqYy4K8+WfV9Tn9k5xHCgo564P9/LfLQcpd9T+me7rkkrfOGxJt1BpjSQ0+038Cjd6O5KIiIiIiIhIjVCxqQ4yGQa9GocxY2AjujcMZeFPedz2wV6+yTzq7WgXFKcllPykiVQENCbk4CICj6wADW0UERERERGROk7FpjosxM/CnZfE889e9fExGzy+JoOn1mZwuMTu7WgXDJfJl4L4MZSGtCPQtorgg/PB5fB2LBEREREREZFzRsWmC0CL2ACev7Iho1tHsymrmL8u3cuSbXlUOtXLpkYYZoqih3I04gr8i7YQlvUGRmWZt1OJiIiIiIiInBMqNl0grGaD4S0ieal/Q5rH+POfzQeZ/PE+fj5c6u1oFwbDoCTiMgpjRmAt3Ut45ixM9nxvpxIRERERERE561RsusDEBfvwUI8kpnRNoKCskimf/MKsDTkcraj0drQLQlnIxeQnjMPkyCc8YyaW8ixvRxIRERERERE5q1RsugAZhkHn+iG8PLAhA5qG88mufG5duofP9hbg0gTW55w9oAm2xJvBMAjLeAWf4p+9HUlERERERETkrFGx6QIWYDVzY9tYpvVtQEyglX+ty+bhVQfILKzwdrQ6r9I3DlvSrVRaIwnN/i9+hRu9HUlERERERETkrFCxSWgU4cfTvS9iYvtYdh0p445le5n33SEqKp3ejlanOS0h5CdNpCKgMSEHFxF4ZDmoZ5mIiIiIiIic51RsEgDMJoMrU8KZMbARnesF83/fH+HOZXvZml3s7Wh1msvkS0H8GEpD2hNoW03wwfngcng7loiIiIiIiMifpmKTVBPub2HypQk8dlk9XMAjqw4w7cssbKUqgJwzhpmi6CEcjeiNf9EWwrLmYFRqlUARERERERE5P6nYJMfVJj6QF/s3ZGTLSNbtL+KvS/fw0Q4blU4N8zonDIOSiJ4UxF6NtfQXwjNfwWTP93YqERERERERkdOmYpOckI/ZxLWtonmhfwMaR/gxa2MuU5b/wp68Mm9Hq7PKg9PJTxiHyVFAeMYMLOVZ3o4kIiIiIiIiclpUbJJTSgrxZerl9ZjUOZ6DxXYmf7yP1zblUmKv9Ha0Oske0Bhb4s1gmAjLeAWf4p+9HUlERERERETEYyo2iUcMw6BHw1BmDGhE7yZhfLDdxm1L9/LV/iJcWkHtrKv0jcWWdCuV1khCs/+LX8FGb0cSERERERER8YiKTXJagnzN3NIhjqf7XESIn5mnPs/kH2syyD1a4e1odY7TEkJ+0kQqApoQcmgRgUeWgwp7IiIiIiIiUsup2CR/SmqUP9P6NmD8xTH8cLCE2z7Yy4Ifj2CvVDHkbHKZfCmIv57SkPYE2lYTcvB/4NLKgCIiIiIiIlJ7qdgkf5rZZDAoLYJ/D2hE24RA3tp6iEkf7eXHgyXejla3GGaKoodwNKI3fkVbCcuag1FZ6u1UIiIiIiIiIselYpOcsehAK/d1S+Lv3ZModzh5YMV+Xvwqm8Iy9cA5awyDkoieFMRejbX0F8IzX8Fkz/d2KhEREREREZFjqNgkZ037pCBeGtCIoc0iWLO3gFs/2MvK3fk4Nc/QWVMenE5+wjhMjgLCM2ZgKc/ydiQRERERERGRalRskrPKz2JiTHoM069sSFKIDy99ncODK/azP7/c29HqDHtAY2yJN4NhJizjFXyKf/Z2JBERERERERE3FZvknLgozJcnrqjP7Z3iOFBQzl0f7uW/Ww5S7nB6O1qdUOkbiy3pFip9ogjN/i9+BRu8HUlEREREREQEULFJziGTYdCrcRgzBjaie8NQFv6Ux20f7OWbzKPejlYnOC0h5CdOoCKgCSGH3iPwyCegIYsiIiIiIiLiZZaauElFRQWPPPIIDoeDyspKOnXqxNVXX83Ro0eZPn06hw4dIjo6mkmTJhEUFATAe++9x6pVqzCZTIwbN442bdoAsGfPHl5++WUqKipIT09n3LhxGIZRE48hf1KIn4U7L4nn8kahzNyQw+NrMrikXhA3toslKsDq7XjnNZfJl4L46wk+tIRA2xrM9nwKY4eBUSNNW0REREREROQYNdKzyWq18sgjj/Dss8/yzDPPsHXrVnbs2MHixYtp2bIlL774Ii1btmTx4sUAZGRksG7dOv71r3/x4IMP8vrrr+N0Vg2/evXVV5k4cSIvvvgiOTk5bN26tSYeQc6CFrEBPH9lQ0a3jmZTVjF/XbqX97fnUelUb5wzYpgpih7C0Yje+B3dSljWHIzKUm+nEhERERERkQtUjRSbDMPAz88PgMrKSiorKzEMg40bN9K9e3cAunfvzsaNGwHYuHEjnTt3xmq1EhMTQ1xcHLt27cJms1FaWkpKSgqGYdCtWzf3OXJ+sJoNhreI5KX+DWke48/rmw4y+eN97Dis4sgZMQxKInpSEHs11tJfCM+chclu83YqERERERERuQDV2Fgbp9PJlClTyMnJoU+fPiQnJ1NQUEB4eDgA4eHhFBYWApCXl0dycrL73IiICPLy8jCbzURGRrq3R0ZGkpeXd9z7rVy5kpUrVwLw1FNPERUVda4ercZYLJY68RwAUVHwQoN41uw6wvOf7eHeT35hcMs4JnZpQLCvhoD9aVFX4Iqoh3nHDCKzXsHV9HYIvMjbqWpcXWorIueS2oqIZ9RWRDyjtiLimQuhrdTYt3qTycSzzz5LcXExzz33HPv37z/hsa4TTHJ8ou3H06tXL3r16uV+ffjwYc/D1lJRUVF14jl+r2U4vNT/It759jBLfshh9c5DjL84hm4NQjQX158WhTlhImHZb2D88AyFcddREZjq7VA1qi62FZFzQW1FxDNqKyKeUVsR8UxdaisJCQnH3V7jq9EFBgbSrFkztm7dSmhoKDZb1VAfm81GSEgIUNVj6ciRI+5z8vLyiIiIOGb7kSNHiIiIqNkHkLMuwGrmxnaxPNe3AdGBVv61LptHVh0gq7DC29HOW5W+sdiSbqHSJ4rQ7P/iV7DB25FERERERETkAlEjxabCwkKKi4uBqpXpvv/+exITE2nXrh2fffYZAJ999hnt27cHoF27dqxbtw673c7BgwfJzs6mSZMmhIeH4+/vz44dO3C5XKxdu5Z27drVxCNIDWgc4cfTvS9iYvtYdh4p4/Zle/m/7w5TUen0drTzktMSQn7iBCoCmhBy6D0Cj3wCp9E7UEREREREROTP8HgYXUZGBkFBQYSFhVFWVsb777+PyWRi4MCB+Pr6nvRcm83Gyy+/jNPpxOVycckll9C2bVtSUlKYPn06q1atIioqirvvvhuAevXqcckll3D33XdjMpm44YYbMJmq6mI33ngjM2bMoKKigjZt2pCenn4Gjy+1jdlkcGVKOJ3qBTNn00HmfX+Yz/YVMLF9HG3iA70d77zjMvlSEH89wYfeJ9C2BrM9n8LYYWBoXiwRERERERE5NwyXhxMh3XPPPUyaNImEhARmz55NdnY2VquV4OBgbr/99nOd84xlZWV5O8IZq0vjOj21NbuYWRtzyC6y061BCOMvjiHcX4WS0+ZyEWD7jKC8T6jwb0RB3ChcZn9vpzpnLsS2IvJnqK2IeEZtRcQzaisinqlLbeWM52w6dOgQCQkJuFwuNm7cyKRJk7j77rv59ttvz1pIkT9qEx/Ii/0bMrJlJOv2F/HXpXv4aIcNp4aDnR7DoCSiBwWx12At/YXwjFmY7DZvpxIREREREZE6yONik9VqpbS0lF27dhEZGUlISAhWqxW73X4u84ngYzZxbatoXujfgMYRfszamMuUT35hT16Zt6Odd8qD25CfMB5TZSHhGTOxlGV6O5KIiIiIiIjUMR4Xm7p06cLUqVN5+eWX6dGjBwB79+4lJibmXGUTqSYpxJepl9djUud4covtTP54H69vyqXEXuntaOcVe0AjbEk3g2EmLHM2PsXbvR1JRERERERE6hCPJ78ZO3Ys3377LWazmRYtWgBgGAZjxow5Z+FE/sgwDHo0DKVdQhBvfXuIpdttfPlLETe1i6VTvSAMw/B2xPNCpU8stqRbCM1+k9DstyiKHkRZaAdvxxIREREREZE64LRmWm7dunW1140bNz6rYUQ8FeRr5pYOcVzWKJSZG3J46vNM2iUEMqF9LLFBPt6Od15wWkLIT5xASM48Qg69h9lhozjiCjA87vAoIiIiIiIicoyTFpsefvhhj3qKPPbYY2ctkMjpSI3yZ1rfBnzws413vjvEbR/s5ZqWUQxqGoHVrF5Op+Iy+VIQP5rgQ+8TaFuD2W6jMHY4GFrxT0RERERERP6ck36jvOyyy9x/z83NZfXq1XTv3p3o6GgOHz7MZ599Rs+ePc95SJGTMZsMBqVF0Ll+MK9tyuWtrYf4bG8BN3eIo3lMgLfj1X6GmaLowVRawwk68gkmRyEF8aNxmf29nUxERERERETOQyctNv02ETjAgw8+yIMPPki9evXc2y699FJmzpzJ1Vdffc4CingqOtDK/d2S2JBRxKvf5PLAiv30ahzKmDbRhPipp85JGQYl4T2otIQRkruA8IxZ5CeMxWkN93YyEREREREROc94PDlLRkYGsbGx1bbFxMSQmaml06V26ZAUzEsDGjG0WQSr9xRw6wd7Wbk7H5fL5e1otV55cBvyE8djqiwkPGMmljK1bxERERERETk9HhebmjVrxowZM8jOzqaiooKsrCxmzpxJ06ZNz2U+kT/Fz2JiTHoM069sSFKIDy99ncMDK/azP7/c29FqPbt/I2xJt4BhJixzNj7F270dSURERERERM4jhsvD7h5Hjx7ltddeY/369TidTkwmEx07dmT8+PGEhISc65xnLCsry9sRzlhUVBSHDx/2dozzjtPl4tPdBby55SAldidDmkVydYtIfC1ade1kTI4iQrPfwFKeTVH0YMpCO3g7ksfUVkQ8o7Yi4hm1FRHPqK2IeKYutZWEhITjbvd4IpugoCDuuusunE4nhYWFhISEYDLpy7rUfibD4IomYXRICuKNLYdY8OMR1u4rZGL7WNolBnk7Xq3ltASTnziBkJx5hBx6D7PDRnHEFWCo3YuIiIiIiMiJndasySUlJWRlZVFWVlZte4sWLc5qKJFzIdTPwp2XxHN5o1Bmbsjh8TUZXFIvmBvbxRAVYPV2vFrJZfKlIH40wYfeJ9C2BrPdRmHscDA04bqIiIiIiIgcn8ffGNesWcPrr7+On58fPj4+7u2GYfDvf//7nIQTORdaxAbw/JUNWbItj3d/OMyWpcX8pXUU/VPCMZsMb8erfQwzRdGDqbRGEHTkY0yOQgriR+My+3s7mYiIiIiIiNRCHheb5s2bx9133016evq5zCNSI6xmg+EtIrn0omBmf5PL65sOsmpPAbd2iCMlSkWUYxgGJeHdqbSEEpK7gPCMWeQnjMVpDfd2MhEREREREallPJ58xel00rp163OZRaTGxQX78FCPJO7tmkBBWSX3fvILszbkcLSi0tvRaqXy4DbkJ47HVFlIeMYMLGWZ3o4kIiIiIiIitYzHxaZBgwaxcOFCnE7nucwjUuMMw6BL/RBeHtiQAanhfLIrn78u3cPafYV4uFjjBcXu3whb0i1gWAjPfAWf4u3ejiQiIiIiIiK1iMfD6JYtW0Z+fj7vv/8+QUHVV/CaOXPmWQ8mUtMCrGZubBdLz0ahzFifw7Qvs1i5O4Cb28eREOJz6gtcQCp9YrAl3Upo9puEZv+XouhBlIV29HYsERERERERqQU8Ljbdfvvt5zKHSK3ROMKPZ/pcxCe78nlr6yHuWLaX4c0jGdY8AqvZ486AdZ7TEkx+4k2E5PwfIYcWY7bbKI7sDYbeIxERERERkQuZx8WmZs2ancscIrWK2WRwZUo4neoF859Nucz7/jCf7Svk5g6xtI4L9Ha8WsNl8qUgfhRBh5YSmP8ZZoeNwtgRYHj8q0VERERERETqGI+/ETocDhYtWsTatWux2WyEh4fTrVs3hg4disWiL5ZSN0X4W/jbpYlc3riYWRtyePjTA3RrEMINF8cQ5q/PPQCGmaPRg3Bawwk68jEmRxEF8aNxmbWqn4iIiIiIyIXI42/Lc+fOZffu3dx0001ER0dz6NAhFi5cSElJCWPHjj2HEUW8Lz0+kJcGNGTBj0dY+GMemzKPMrpNNH2SwzAZhrfjeZ9hUBLenUpLGCG58wnPmEV+wlic1nBvJxMREREREZEa5vHkKl9//TX33nsvrVu3JiEhgdatW/O3v/2Nr7766lzmE6k1fMwmrmsVzQv9G9A4wo9ZG3OZ8skv7Mkr83a0WqM8uDX5iTdgqiwkPGMGlrIMb0cSERERERGRGuZxsUlLwItUSQrxZerl9ZjUOZ7cYjuTP97H65tyKbFXejtarWD3b4gt6RYwrIRnzsaneJu3I4mIiIiIiEgN8rjYdMkll/D000+zdetWMjIy2Lp1K88++yyXXHLJucwnUisZhkGPhqHMGNCIKxqH8f52G7ct3ctX+4tUmAUqfWKwJd2CwyeG0Oy38CtY7+1IIiIiIiIiUkMMl4ffjB0OBwsXLuSLL77AZrMRERFB586dGTZsGFar9VznPGNZWVnejnDGoqKiOHz4sLdjyHFsP1TKzA057Msvp31iIDe1iyU2yMfbsbzPWUFozjx8S7ZTHNad4sjeYHhc4/7T1FZEPKO2IuIZtRURz6itiHimLrWVhISE4273uNh0vlOxSc61SqeLD3628c53h3C6YGTLKAalRWAxXeATiLsqCTq0lIDC9ZQFtaIwdgQY53YlP7UVEc+orYh4Rm1FxDNqKyKeqUtt5UTFJo+7GCxevJhdu3ZV27Zr1y6WLFlyZslE6gizyWBQWgT/HtCIixMC+e/WQ0z6cC8/HSzxdjTvMswcjR7E0ci++B39jrDM/2BUXuDviYiIiIiISB3mcbHpww8/JCkpqdq2pKQkPvzww7MeSuR8Fh1o5f5uSTzYPZFSu5P7V+znpa+zKSxzeDua9xgGJeHdKYgdibVsP+EZszDZ87ydSkRERERERM4Bj4tNDocDi6X60BeLxUJFRcVZDyVSF3RICubfAxsxtFkEq/cUcOsHe/l0d/4FPYF4eXBr8hNvwFR5lIiMmVjKMrwdSURERERERM4yj4tNjRo14pNPPqm2bfny5TRq1OishxKpK/wsJsakx/Cvfg1ICvHhxa9zeHDlfvYXlHs7mtfY/RtiS7oZl2ElPHM2PsXbvB1JREREREREziKPJwg/cOAA//jHPwgLCyM2Npbc3Fzy8/N56KGHjhleVxtpgnDxNqfLxae7C3hzy0FK7E6GNIvk6haR+FrO/epstZHhKCIs+00s5Vkcjb6K0tBOZ+3aaisinlFbEfGM2oqIZ9RWRDxTl9rKWVmNrqysjE2bNnHkyBEiIyNp27Ytfn5+Zy3kuaRik9QWBWUO3thyiFV7CogJtDKxfSztEoO8Hcs7nBWE5szDt2Q7xWHdKI7sA8aZF9/UVkQ8o7Yi4hm1FRHPqK2IeKYutZUzXo0OwM/Pj9TUVJo2bUqXLl3Om0KTSG0S6mfhzkvi+Wev+viYDR5fk8FTazM5XGL3drSaZ/KhIH4UJSEdCcxfS0juu+C8AN8HERERERGROsRy6kOqHD58mBdeeIF9+/YB8NZbb/H111+zdetWbr755nOVT6TOahEbwPNXNmTJtjze/eEwW5YW85fWUfRPCcdsMrwdr+YYZo5GD8JpjSDoyEeYHIUUxI/GZQ7wdjIRERERERH5Ezzu2TR79mzS09N588033avStWrViu++++6chROp66xmg+EtInmpf0OaRfvz+qaD/O3jfew4XOrtaDXLMCgJ70ZB7LVYyw4QnjELkz3P26lERERERETkT/C42LRr1y4GDx6MyfT/TwkICKCkpOScBBO5kMQF+/BwzyTu7ZpAflkl937yC7M25HC0otLb0WpUeXAr8hNvwFR5lIiMmVjKMrwdSURERERERE6Tx8Wm0NBQcnJyqm3LyMggKirqrIcSuRAZhkGX+iG8PLAhA1LD+WRXPn9duoe1+wo5jXn8z3t2/4bYkm7GZVgJz5yNT/E2b0cSERERERGR0+BxsWngwIE8/fTTrF69GqfTyRdffMH06dMZNGjQucwncsEJsJq5sV0sz/VtQFSAlWlfZvHoqgNkFVZ4O1qNqfSJIS/pFhw+MYRmv4V/wVfejiQiIiIiIiIe8rjYdNlll/GXv/yFr7/+msjISNauXcvIkSPp2rXrucwncsFqHOHHM30uYkK7WHYcKeOOZXv5v+8PY690ejtajXBZgrElTqAioCnBh94n8PBH4Lownl1EREREROR8dspi0549e9i/fz8AHTp04NZbb+Wiiy4iLy+PLVu2UFZWds5DilyozCaD/qnhvDywEZ3qBTHvu8PcsWwf3+YUeztazTD5UBA/ipLQTgTmryUk911w2r2dSkRERERERE7ilMWmN954g/z8fPfrV155hZycHHr16sWBAweYO3fuucwnIkCEv4W/XZrIo5fVw+ly8fCnB/jXl1nklzq8He3cM0wcjbqKosh++B39jrCs/2BUamECERERERGR2uqUxabMzEzS0tIAKC4uZsuWLdx+++307duXO++8k02bNp3zkCJSJT0+kBf7N+SalpF8ub+IW5fu4aMdNpx1fQJxw6A0vBsFsddiLTtAeMYsTPY8b6cSERERERGR4zhlsamyshKLxQLAzp07CQsLIyEhAYCoqCiKiy+Q4TwitYSvxcR1raJ5oX8DGkf4MWtjLlM++YU9eXV/SGt5cCvyE2/AVHmUiIyZWMoOeDuSiIiIiIiI/MEpi0316tXjq6+qVoL68ssvadmypXtfXl4eAQEB5y6diJxQUogvUy+vx6TO8eQW25n88T5e35RLib3S29HOKbt/Q2xJt+AyrIRnvopP8U/ejiQiIiIiIiK/c8pi01/+8hdeffVVxo0bx+bNmxk8eLB737p160hNTT2X+UTkJAzDoEfDUGYMaMQVjcN4f7uN2z7Yy1cHinDV4aF1lT7R5CXdgsMnltDsufjnf+XtSCIiIiIiIvIrw+XBN9LS0lKys7OJj4/H39/fvT0rKws/Pz8iIiLOacizISsry9sRzlhUVBSHDx/2dgypxbYfKmXmhhz25ZfTPjGQm9rFEhvk4+1Y546zgtCc/8O3ZBvFYd0ojuwDhkltRcRDaisinlFbEfGM2oqIZ+pSW/ltmqU/OmXPJgB/f38aNWpUrdD020XPh0KTyIWiabQ//+rXgPEXx/B9bgm3fbCXhT8eweGso72cTD4UxI+iJLQTgflrCcn9P3DavZ1KRERERETkgmbxdgARObvMJoNBaRF0rh/Mq9/k8t+th1izt4BbOsTRLKYOzrFmmDgadRWVlgiCj3yIyVEE4Xd6O5WIiIiIiMgFq0aKTYcPH+bll18mPz8fwzDo1asXV155JUePHmX69OkcOnSI6OhoJk2aRFBQEADvvfceq1atwmQyMW7cONq0aQPAnj17ePnll6moqCA9PZ1x48ZhGEZNPIbIeSU60MoD3ZNYn1HEqxtzuX/Ffno1DmVMegwhvmZvxzu7DIPS8K44LaGEHJwPP/yDwIA2lAc2xeGbAIZHnThFRERERETkLKiRYpPZbGb06NE0atSI0tJS7rvvPlq1asWaNWto2bIlgwcPZvHixSxevJhRo0aRkZHBunXr+Ne//oXNZuPxxx/nhRdewGQy8eqrrzJx4kSSk5N58skn2bp1K+np6TXxGCLnpY5JwbSOC+Td7w+zZFse6zOOMi49mssahda5Qm15cCvyLSGEFX5KgG0VgbZPqTQHUxGYSnlAU+wBTXCZfL0dU0REREREpE6rkf/dHx4eTqNGjYCq+Z8SExPJy8tj48aNdO/eHYDu3buzceNGADZu3Ejnzp2xWq3ExMQQFxfHrl27sNlslJaWkpKSgmEYdOvWzX2OiJyYn8XEmPQY/tWvAUkhPrz4dQ4PrtzP/oJyb0c76+z+DXA1n8Lhhg9SGDMCu38DfI9+T1jOXKL2PE5o1n/wz1+HyW7zdlQREREREZE6qcbnbDp48CB79+6lSZMmFBQUEB4eDlQVpAoLCwHIy8sjOTnZfU5ERAR5eXmYzWYiIyPd2yMjI8nLy6vZBxA5jzUI9+OJK+rz6e4C3txykLuW7WVIs0iubhGJr6VuDTVzmQMpC7mYspCLwVWJtXQfviXb8SneTvDhpQQfXorDJ5bygKZUBDbF7ldfw+1ERERERETOghotNpWVlTFt2jTGjh1LQMCJJyp2uY6/ctaJth/PypUrWblyJQBPPfUUUVFRpxe2FrJYLHXiOcT7ro2Opm+ri5jxxV4W/HiQLw8c5e4ejencsG6sLnn8thILdATAWZoD+d9jtn1HQMEXBOZ/hssSCGEtcYW1grDmYKmDk6mL/IH+XRHxjNqKiGfUVkQ8cyG0lRorNjkcDqZNm0bXrl3p2LHqC19oaCg2m43w8HBsNhshISFAVY+lI0eOuM/Ny8sjIiLimO1HjhwhIuL4X4579epFr1693K8PHz58Lh6rRkVFRdWJ55DaY+LFEXRJ9GPmhhzuef8nOtcP5sa2MUQGWL0d7Yycuq1YwJoOMekYkWX4lO7Ep3gbvnnfYzr8NS5M2P0buHs9VVqjoI7NbyUC+ndFxFNqKyKeUVsR8UxdaisJCQnH3V4jY0ZcLhezZs0iMTGRAQMGuLe3a9eOzz77DIDPPvuM9u3bu7evW7cOu93OwYMHyc7OpkmTJoSHh+Pv78+OHTtwuVysXbuWdu3a1cQjiNRZLWIDeP7KhoxqHcU3mUe5delelm7Po9LpeU/C85nL7Ed5UEuKYq/mcMMHyEu8mZLwbpgqSwg+8iGR+/9FxP5pBB36AGvJLnA5vB1ZRERERESkVjNcpzM27U/avn07Dz/8MPXr13evfnXttdeSnJzM9OnTOXz4MFFRUdx9990EBQUBsGjRIlavXo3JZGLs2LHuFed2797NjBkzqKiooE2bNowfP96jFbWysrLO3QPWkLpU/ZTaKaeoglc25rI5u5hG4b7c0iGOlCh/b8c6bWerrZjsNvc8Tz6lezBcDpwmXyoCUqgIaEp5YAouc9BZSCziHfp3RcQzaisinlFbEfFMXWorJ+rZVCPFptpAxSYRz7hcLtYdKOK1bw5iK3XQNzmMUW2iCfIxezuax85JW3FW4FO6C9/iquKTubIIFwYOv3qUBzSlPLAplT5xGm4n5xX9uyLiGbUVEc+orYh4pi61lRMVm2p8NToRqd0Mw6BL/RDS4wN5+9vDfLjDxtcHihjfNpauFwV71JOwTjL5UBHYjIrAZuByYinPxqdkO77F2wnKW05Q3nIqLaHueZ4q/BuD6fye+0pEREREROTPULFJRI4rwGrmpnax9GwYyswNOUz7MotPdwcwsX0cCSE+3o7nXYYJh18iDr9ESiIux+QoxKdkBz7F2/Er2kJA4XpchpUK/yaUBzalIjAVpyXU26lFRERERERqhIpNInJSTSL9eKbPRXy8M5+53x7ijmV7Gd4ikmHNIrCaa2SNgVrPaQmhLKQdZSHtwGnHp3Svu9eTb8k2OAR234Rf53lqisM3EQy9dyIiIiIiUjep2CQip2Q2GfRPDeeS+sH8Z1Mu8747zGd7C7m5Qyyt4wK9Ha92MVmpCEyhIjCFo1EDMVccdE8yHmBbTaBtFZXmIHfhyR7QBJfJ19upRUREREREzhoVm0TEYxH+Fv52aSKXNy5m1oYcHv70AN0bhDD+4hjC/PXr5BiGQaVvLCW+sZSEd8eoLManZEdVj6fiH/Av+gYXZir8G1ERWFV8clojvJ1aRERERETkjOjboYictvT4QF7s35AFPx5h0U9H+CbrKNe3iaZ3kzBMF+oE4h5wmQMpD06nPDgdXJVYy37Bp7hquF3w4aUEH16KwyfGPcm43a8+GOfPKoAiIiIiIiKgYpOI/Em+FhN/aR1N9wYhzNqYy8wNuazaU8AtHeJoGO7n7Xi1n2HG7t8Iu38jiqOuxFxx2D3PU0D+FwTmr8Vp8qciIKVqkvGAVFxmf2+nFhEREREROSUVm0TkjCSF+vL45fX4bF8h/9l0kLs/2sfA1HCubRWNv1WTYHuq0ieKUp9LKQ27FKOyDJ/Snb/2evoZv6Pf4sKE3e+iX1e3a0qlNRrUi0xERERERGohFZtE5IwZhkGPhqG0Swjiv1sPsWS7jS/2F3FTu1g6JQVhqChyWlxmP8qDWlIe1JIilxNLeQa+xVWTjAcf+QiOfITDGvHrJONp2P0bgKFf5yIiIiIiUjvo24mInDVBvmZu7RjHZY1Cmbkhh6fWZtI+MZAJ7eKICbJ6O975yTDh8KuPw68+xZG9Mdnz8Sn5Gd/ibfgXbiCgYB1Ow5eKgOSqScYDUnFZgrydWkRERERELmAqNonIWdc02p9/9WvA0p/zmPfdYW77YA8jW0ZxVVoEFpN6OZ0JpzWMstCOlIV2BGcFPqW73b2e/Ip/wIWBwzfJPdzO4ROv4XYiIiIiIlKjVGwSkXPCbDIYnBZJl/ohvPpNLm9uPcTqvVUTiDeLCfB2vLrB5ENFYBoVgWngcmGpyManeBu+xdsJylsBeSuotIS6V7er8G8MJvUwExERERGRc0vFJhE5p6IDrTzQPYn1GUW8ujGX+1fsp1fjUMakxxDia/Z2vLrDMHD4JuDwTaAk4nJMjiJ8Sn6u6vFUtIWAwvW4DCsV/o3dvZ6cllBvpxYRERERkTpIxSYRqREdk4JpHRfIu98fZsm2PDZkHGVsejSXNQrVBOLngNMSTFlIO8pC2oHLgbV0L77F2/Et3oZvyXY4BHaf+Kp5ngKb4vBNAkOrB4qIiIiIyJlTsUlEaoyfxcSY9Bi6Nwhh5oZcXvw6h0/3VA2tqxfq6+14dZdhwR6QjD0gmaNRAzDbD7nneQqwrSHQthqnOYjygNSq4XYBTXCZ/LydWkREREREzlMqNolIjWsQ7seTvevz6e4C3txykLs+3MvgtEiubhGJr0W9a84pw6DSJ4YSnxhKwrthVJbgU7Lj115PP+JftAkXZuz+Dd3D7Sqtkd5OLSIiIiIi5xEVm0TEK0yGwRVNwuiQFMQbWw6y4McjfP5LIRPbxdI2Mcjb8S4YLnMA5cFtKA9uA65KrGX78Snejm/xdoIPfwCHP8BhjXEXnux+9cHQXFsiIiIiInJiKjaJiFeF+lm485IELmsUyqwNuUxdk0Hn+sHc2DaGyACtnFajjKoeTXb/hhRH9cNsP+IuPAXkf0lg/lqcJj8qAlKrik8BKbjMWllQRERERESqU7FJRGqFlrGBPH9lQxZvO8L/fjjC5qxiRrWO4sqUcMwmTSDuDZXWSErDulAa1gXDWY5Pyc6q4lPJz/gd/RYXBna/i9yTjFdaY0CTvYuIiIiIXPBUbBKRWsNqNhjRIoquF4XwysZcXtt0kNV7qyYQT47093a8C5rL5Et5UAvKg1pQ5HJiKc9wTzIedORjgo58TKUl3D3crsK/ERj6J0ZERERE5EKkbwIiUuvEBfvwcM8k1h0o4rVvDnLPx7/QLyWMUa2jCfTRfEFeZ5hw+NXH4Vef4sjemBwF7uF2/oUbCSj4CqfhQ0VAclWvp4BUXJZgb6cWEREREZEaomKTiNRKhmHQpX4I6fGBvP3tYT7cYeOr/UWMbxtL14uCMTRcq9ZwWkIpC+1IWWhHcFbgU7rn1+F22/Er/hEAu2/Sr72e0nD4xGu4nYiIiIhIHaZik4jUagFWMze1i6Vnw1Bmbshh2pdZfLo7gJs7xBEf7OPtePJHJp+qYXSBTTnqcmGpyPm119M2AvM+JShvJZXmEPc8TxX+jcGkn6OIiIiISF2iYpOInBeaRPrxTJ+L+HhnPnO/PcTtH+xleItIhjWLwGo2eTueHI9h4PCNx+EbT0lETwzHUXxLfq4qPhVtxb9wAy7DQoV/41+H2zXFaQ3zdmoRERERETlDKjaJyHnDbDLonxrOJfWD+c+mXOZ9d5jP9hZyS4dYWsUFejuenILLEkRZSFvKQtqCy4G1dB++xdvw/XWFu2CWYPeJd/d6cvgmgaFCooiIiIjI+UbFJhE570T4W/jbpYlc1ugor2zM5aFPD9CjQQjjLo4hzF+/1s4LhgV7QBPsAU04GjUAs/2Qe5LxANtnBNpW4zQHUh6QWjUsLyAZl8nP26lFRERERMQD+lYmIuetixOCeLF/AAt+PMKin46wMeso17eJ5rrISG9Hk9NhGFT6xFDqE0NpeDeMylJ8SnZU9Xgq3oZ/0WZcmLD7N6ya5ymgKZU+Ud5OLSIiIiIiJ6Bik4ic13wtJv7SOpruDUKYtTGXmRtyWbn3KC2jfWkS6UdyhD/RgRatXncecZn9KQ9uTXlwa3BVYi074F7dLvjwMmAZDmu0e54nu/9FYJi9HVtERERERH6lYpOI1AlJob48fnk91uwt5OPdRby/PQ+Hs2pfiK+ZJhF+NIms+pMc6U+EhtudHwwzdv8G2P0bUExfTPY8fIu341O8Hf/8dQTkf47T5EdFQMqvvZ5ScJk1f5eIiIiIiDfp25aI1BmGYdCzUSgjOjQmO/cg+/LL2XWkjF15Zew6UsaCH4/gdFUdG+Fv+bXn069FqAg/Qvz0K7G2c1ojKA3rTGlYZwxnOdaSXVXFp5Lt+B39DhcGdr+L3L2eKn1iQL3aRERERERqlL5ZiUidZDWbSI70JznS372t3OFkj62q8PRbEWpjxlF+rT8RE2gl+dfCU5NIPxpH+BHoo+FZtZXL5EtFUHMqgpqDy4mlPNNdeAo68jFBRz6m0hLunuepwr8hmKzeji0iIiIiUuep2CQiFwxfi4m06ADSogPc20rslezOK2Pn7wpQX+4vcu9PCPapKkD9WoRqFOGHn8XkjfhyMoYJh189HH71KI68ApOjAJ/in/Et2Y5/4TcEFHyF0/DBHtDk1+JTKk5LiLdTi4iIiIjUSSo2icgFLcBqpmVsIC1j//88P4XlvxWgStl1pIwfckv4bF8hACYD6oX4uud/ahLhR8NwX6xmFaBqE6cllLLQDpSFdgCnHZ/S3e5Jxn2LfwLA7pv4a+EpDYdvgobbiYiIiIicJSo2iYj8QYivmfT4QNLj/38BKq/Uwe4jZezMqypAfZN5lE/3FABgMcFFYb40ifD/dQJyP+qF+mIxqXhRK5isVAQ2pSKwKUddLswVOfj+WngKzFtFUN6nVJpDqAhMrSo++TcBk4+3U4uIiIiInLdUbBIR8UCEv4WIpCDaJwUB4HK5OFziYNeRX3tA5ZXxxf5CPtmVD4CP2aBhuC9NIv1pElFVgEoI9sGsApR3GQaVvvGU+MZTEtETo/IovsU78CnZjm/Rd/gXbsRlWKjwb+SeZNxpDfd2ahERERGR84qKTSIif4JhGEQHWokOtHJJ/WCgqgCVc9T+6/xPVQWoT3fns+znqinI/SwmmkRUFaAa/1qAiguyYmj4lte4zEGUhVxMWcjF4HJgLf0F3+Jt+JRsJ/jQ+wTzPg6fOMp/LTw5/OqBoSGTIiIiIiIno2KTiMhZYhgG8cE+xAf70K1B1eTTlU4XmUUVv66AV8rOI2Us+9mG3VlVgAryMf26+p2/exW8qACLClDeYFiwBzTGHtAYGIC54pB7nqcA21oCbWtwmgIoD0ytWt0uIAWX2c/bqUVEREREah0Vm0REziGzyaB+qC/1Q325rFEoAA6ni/355ezKK3MPw3vvpyNUVtWfCPUzk/xr4Sn51yJUmL9+Xde0Sp9oSn2iKQ3vilFZik/JTnxLtuFbvB3/oi24MGH3b0B5QFMqAtOo9InydmQRERERkVpB315ERGqYxWTQKMKPRhF+9G5Sta2i0sk+W3nVELy8ql5Qm7OL+bUDFJEBFpJ/Xf0u+ddheMG+Zu89xAXGZfanPLgV5cGtwOXEWrbf3esp+MiHcORDHNYo9zxPdv8GYOjnIyIiIiIXJhWbRERqAR+ziZQof1Ki/N3bSu1O9tiqej/9VoD6+sBR9/64ICtNfleAahThS4BVBY5zzqjq0WT3b0AxfTHZbfgWb8enZDv++V8RkP8FTpMvFQEpVAQ0pTwwFZc58NTXFRERERGpI1RsEhGppfytJprHBNA8JsC97WhFJbvdw+/K2HG4lC9+KQLAABJDfKoVoBqG++Jr0YTW55LTGk5p2CWUhl2C4SzHWrIb35Lt+BRvx+/o97gwsPvV/7Xw1JRKn1jQnFwiIiIiUoep2CQich4J8jHTOi6Q1nH/v6dMQZnD3ftp55Eyvs0uZs3eQgBMBlwU5kvjiP9fgLoozBerWcWOc8Fl8qUiqBkVQc3A5cRSnuXu9RSU9wlBeZ9QaQmjPLBp1STj/o3AZPV2bBERERGRs0rFJhGR81yon4W2iUG0TQxybztSYq9WgFqfcZSVuwuAqjmjGob7ule/axLhR71QX8wmFaDOKsOEwy8Jh18SxZG9MDkK8Sn+Gd+S7fgXbiKg4GtchpWKgORfJxlPxWkJ8XZqEREREZEzpmKTiEgdFBlgJTLASsd6wQC4XC4OFlcvQH22r5CPduYD4GuumrTcXYCK9CMh2AeThnudNU5LCGWh7SkLbQ9OOz6le/Ap2Y5v8XZ8i3+CQ2D3TXQPt3P4JoChIZAiIiIicv5RsUlE5AJgGAaxQT7EBvnQ5aKq3jNOl4usoorfTUBexie78ln6c9USeAFW0++G31UVoGICrRgqQJ05k5WKwFQqAlM5GnUV5opc9zxPAbZVBNo+pdIcTEVgatXqdgFNcJl8vZ1aRERERMQjKjaJiFygTIZBUogvSSG+9GgYCkCl08WBgnJ38WlXXhlLf7bhcFYVoIJ9zVW9n35XgIoM0JxDZ8QwqPSNo8Q3jpLwHhiVxfgW/7/27jXGrfO+8/jvec45vA3nyhldLdnWxXHcOJVdGdl40zhdK63RBF2h6AYo4BZxm15Q9038Ki16CeAGNdA6TgrYKFAUDQoUSAtsImy72XbjKmsXdtuocVy3sRVLlhVHtmxpyJE0MySHPJd9QQ7vnJs4Qw7n+wEEkuc8PPNwRmdI/fR//uf7laqnhf9Q8sa/KZKjUupQterp/Qq9yX7PGgAAAOiKsAkAUONYo9smE7ptMqEThyvbykGkH1xb0vlcQeeyRb2RK+p/vppVNX/SZNKtLb87Wr0dT/D2slGRM6Li2L0qjt0rRYG8wsVa1dPo7N9qdPZv5cd2V/s83aly4iDL7QAAADBQ+NcAAGBFnmNqfZweOlrZtuSHenOuHkCdzxb1b28vqJo/adeIq8NTyVr10+GphNIxp2+vYdsyjsqpwyqnDkvTn5BTuqpY/vuKL55V6to/aeTacwptsrrc7v0qpY4qcpL9njUAAAB2OMImAMC6xV2rO2eSunOmHmzky4Eu5JoDqH/+4Xxt/75RT0emkrXg6tBkQkmPipz1CGIzKsRmVJj4iExQVKxwTrHF1xRffF2J+ZcVyaqcvK1W9RR40xI9tgAAALDFCJsAAD2R8hx9YHdKH9idqm2bXwr0RrX/07lcQa9ezev5H9yQJFkj3TIWq4RP1RDq9sm4Yg4B1FpETkJL6bu1lL5b81Eot/jD+nK77Dek7Dfke5na1e3Kydskw9s+AAAANt+WfOp85pln9NJLL2l8fFxPPvmkJGlhYUFPPfWUrl69qpmZGX32s59VOp2WJH3961/X6dOnZa3VI488omPHjkmSLly4oKefflqlUkn33HOPHnnkEa6KBAADbDTu6NjeER3bO1Lbdq3g63yuqHPZgs5ni/rOO4s6faESQDlGunUiXun/lEnqyFRCByfici2/61dkrPzkrfKTt2ox81Oy5bla8JS88a9KXX9BoYmrlDqq0sidWhp5nyIn3e9ZAwAAYEhtSdj0sY99TA899JCefvrp2rZTp07p7rvv1smTJ3Xq1CmdOnVKDz/8sC5duqQXX3xRX/ziFzU3N6fHH39cX/7yl2Wt1Z/92Z/p137t13T06FH94R/+oV5++WXdc889W/ESAAA9MpF0dXx/Wsf3V8KOKIo0m/frV8DLFvTiW/P6v+evS5I8a3T7ZHMAtX8sJocAqqvQm1Rh/MMqjH9YCkuKFc4rvlgJnxKL/6lIRn7igEzhg4r7Iwq8jAJvmn5PAAAA6IktCZvuuusuXblypWnbmTNn9PnPf16S9MADD+jzn/+8Hn74YZ05c0b333+/PM/Trl27tGfPHp0/f14zMzMqFAq64447JEkf/ehHdebMGcImANjmjDGaGfE0M+LpwwdGJVUCqHcXypXwKVcJoE5fuKFvvH5NkpRwjQ5NJpoCqL2jHtWundiYSiN3qTRylxRFcpfeUSx/VvHFs9Lb/1vjtbbuUmhHFMQy8r1MLYAKvGkFsYwim+jjiwAAAMB20rfmDdevX9fk5KQkaXJyUjduVJZQ5HI5HT16tDZuampKuVxOjuMok8nUtmcyGeVyua7Hf/bZZ/Xss89Kkp544glNT09vxsvYUq7rDsXrADYb58pwmJmR7r69/jgII/3wWkFn31vQa+/N6+yVBf39uev6X2fnJFWW7L1vV1p37h7VnbvSev/utHaPxgmg2sxI+lFJkjWhgsXLUvGKVLwiU3xPbvGK3OJFmfnvNj0r8kalxG4pPqMouVtK7Ko8TuySHIIoDDfeV4C14VwB1mYnnCsD1yk0iqJ1be/mxIkTOnHiRO3x7OzsTc1rEExPTw/F6wA2G+fK8EpLOj5jdXxmXNK4/DDSD68vVRqQV6ugvvrSdflhZfx43Kld/e7IVKUKajI5cG99fTM9Pa3ZfFzSASl2QIpJGqvuDEtyyjm55Vk55WzlT2lWTv57cmb/uek4gTNarYTK1CqhKlVRU4psfKtfFtBzvK8Aa8O5AqzNMJ0r+/bt67i9b5+4x8fHNTc3p8nJSc3NzWlsrPLpNpPJKJvN1sblcjlNTU21bc9ms5qamtryeQMABodrjW6fTOj2yYQ+fqSyrRSE+sG1pUr4VA2gvns5q7D6fxaZpNsUQB3JJDUWd/r3IgaVjSmI71EQ39O+LyzJLWfllGfllKpBVHlWsfzrcoLvNA2tBFGNAVR9mZ5sbIteDAAAALZS38Km48eP67nnntPJkyf13HPP6b777qtt/5M/+RN98pOf1NzcnC5fvqwjR47IWqtkMqnXX39dR48e1fPPP6+HHnqoX9MHAAyomGN1NJPU0Uy92XXRD/Vmrqhz1Sbk57JF/eulhdr+3WmvGjwlarcpjwCqKxuTH98rP763bZcJl5oroapBVHzxrGyw0DQ2cMYUxKZrFVH1XlEZyXpb9WoAAADQY1sSNn3pS1/Sq6++qvn5ef36r/+6PvWpT+nkyZN66qmndPr0aU1PT+uxxx6TJB04cEAf/vCH9dhjj8laq1/+5V+WtVaS9JnPfEbPPPOMSqWSjh07RnNwAMCaJFyr9+9K6f27UrVti6VAb+Tq1U/nskW98NZ8bf/+sVh16V0lgDo0lVDctf2Y/rYS2bj8+D758faSahMWmyqhnPKs3HJW8YVXZcPF+jFkFLpjLZVQ1eood4ogCgAAYMCZaL3NkLapd955p99TuGnDtK4T2EycK9ioG0W/evW7egCVK/iSJGukA+PxWvh0JJPQbRNxec72DaAG6VwxQaEhhMpWekVVgykb5mvjKkHUeHMlVK06akoy9ORC7w3SuQIMMs4VYG2G6VwZuJ5NAAAMmrGEq3v3pXXvvnRtWzZfrgVQb+SK+valBT37xnVJkmulWycSTQHUwfG4HMsV8NYrcpLynVvkJ25p21cJoqqVUKV6IJVYeEU2LNSPIaPQnahUQzUsz6sHUSyNBAAA2AqETQAArCCT8pRJefrQLaOSKldHvbro61yuUKmAyhb1Txdv6O/PXZMkxZxK0/KjtSvgJbRvLCZrCKA2qhJEHZCfOKClln0mWKxUQpUarppXnlVi/mXZsFg/hqxCd0J+rDGAWr6dJIgCAADoIcImAADWwRijXWlPu9Ke/uvBypVUwyjSu/NlncsWalVQ3zx/TX8XVFaqJ12rw1NxHckkawHU7rQnQwB10yJnRL4zIj9xsGVHJBMuyill61fOq956hbdko3psFckq8CZbAqjq8jx3giAKAABgnQibAAC4SdYY7RuLad9YTA/cPi5JCsJIb98o1QKoc9mi/u77c/LDSgA1GrM6XA2fjmQqAVQm6RJA9Yoxipy0/GRafvLW5n1RJBMsVEOohiCqNCuvcFE2KtWHymkPomIZ+d60QndCMtu3ZxcAAMBmIWwCAGATONbo4ERcByfievBwZVs5iPTD60s6ly3qfK6gc9mivvZqVtX8SRMJp1r5lNSRTCWEmkjwVt1zxihyR1V2R1VO3ta8L4pkg/n6krxSvSIqVrggE5XrQ+Uo8KaaKqH8aiAVuuMEUQAAYMfiEywAAFvEc4wOTSV0aCqhn9KEJGnJD3Xx2lL1CniVAOo77yxq+VKx0ym32v+pGkBNJZSOs6xr0xij0B1T6I6pnLy9eV8tiFq+Ut7ylfOyihXeaA6ijKvAnVJQ7RHlN1RGhe4YQRQAABhqhE0AAPRR3LV633RS75tOSpqUJBXKoS7kirX+T+dyBf3zDxdqz9mT9ioBVCaho1NJ3T4VV8ojgNp0TUHUoeZ9UVgJohoqoZaX6cXy52Qivz7UuLVqKL+xWXkso9AZJYgCAADbHmETAAADJulZ/cjulH5kd6q2bWEpqIRPuaLOZws6e7Wgf/rBvCTJSNo/FqsHUJmkbpuIK+4SWmwZYxW64wrdcZV1uHlfFMr6N5oqoSrVUbOKLX5fRkF9qPGqIVQ9gFoOoypBFD29AADA4CNsAgBgG0jHHR3bO6Jje0dq264V/eryu0oA9d3Li/rWmzckSdZIt07EGxqQJ3VwPC7PIazYcsYq9CYUehMq60jzviiU9a/XK6GqlVFu6Yrii2ebgqjQxOr9oZavllcLotIEUQAAYGAQNgEAsE1NJFwd35/W8f1pSVIURcoWqgFUtqhzuaL+5Yfz+uYb1yVJnjW6bbIeQN0TJuSUfI3FHVmCiv4wVqE3qdCbVFlHm/dFgax/XW5rj6jSZcUXX5VRWBsamlhD+FTvD+V704qcEYIoAACwpQibAAAYEsYYTac8Tac8/ZcDo5IqAdR7C+WG/k9F/b83b+j/nLsm/cu7kiTXSpMJV1MpV1PJ5T9e/XH1dsSzMoQWW8c4Cr0plbwpKdWyLwrk+NeaekQ55azcpbcVX/hecxBl400BVK1heWxakU0RRAEAgJ4jbAIAYIgZY7RnNKY9ozF95NYxSVIYRXrnRklzYVxvXZlTNl9WruArV/B16UZJr7yX12IpbDtWzDG1MCrTFER59ZAq5SpBr6jNZ5xaFVObKJBTnqsFUMu3XvGS4gv/IVO71qEU2kRzJVRDdVTkjLQfGwAAYA0ImwAA2GGsMbplPK5j09Oanew8ZskPKwFU3le24GuuGkZVHlcqpbJ5X6UganvuiGc1mWytlFoOqCrB1GTSpX/UZjFOZUldbLp9X+S3BFFZuaVZecW3FF94pSWIStYroWKNy/OmFTnJLXxBAABguyFsAgAAbeKu1d7RmPaOxrqOiaJI+XJYq4rK5v3a/Vz1/qtX8soVfPnthVIaizu1EGqyIZRaDqkyKU/jcUeOJZTqGeMqiM0oiM2074t8OeVc9Up51SCqPCuveFHxhX9vCaJSbZVQ9SAqsYUvCAAADCLCJgAAsCHGGI3EHI3EHB0Yj3cdF0aR5peCphCqNZy6MLek60VfYUuhlDWVRuhNIVSHqqnRuEM/qZtlXAWxXQpiu6TWFXRhuRpEZetXzivPKla4IGf+u81DnRH53rQCb6oWQAW1HlHd/54AAIDhQdgEAAA2lTVG4wlX4wlXt3dZtidJQRjpWtFvC6WWH19ZKOu1qwXNLwVtz3WtaauM6tRfKunS5HxDrKcgvltBfHf7vrBUC6Lc5eV5pVnF8m/ICZqDqMBJNzcrb+wRRRAFAMDQIGwCAAADwbFGmZSnTMqTOvS9XlYKwnoPqcZgqnr71rUlvXx5Ufly+9q9hGs6X3GvJaSK0+R87WxMQXyPgvgelVr3haVqCJVtWp4Xy78uJ5hvGho4o/UgKtZw1TwvI9nuyzkBAMDgIWwCAADbSsyx2p2OaXd65QCiUOsnVe5YKfV6tqDcpc5NztMx2xJCeW2B1GTSlUs/qZXZmIL4XgXxvW27TLhUa1LulOoNy2P5s3LmF5rGBs5YWyWUX62OkvW26tUAAIA1ImwCAABDKelZ7fdi2j+2cpPzxVJjk/NyWyh16d285gq+WjMpI2ks4XToI1UPpjJJV2MJR5ale20iG5cf3yc/vq9tnwmLtSqo5T5RTjmr+OKrssFi09jAHW9amufX7k8RRAEA0CeETQAAYMcyxigdd5SOOzo4sXKT8xvFoC2IagyozueKul4M1Fon5RhpItncR2o5iJpKebVt6Rj9pJZFNiE/sV9+Yn/bPhMUa+HT8q1bmlV84T9lw3z9GDIKa0FUtRIq1hBEGT4GAwCwWXiXBQAAWIU1RhNJVxNJV4dWGOcvNznP+8p2aHT+7nxZ37uS10KpvZ+UZ00thJrs2Oi8EkwlvZ3dTypyEvKdW+QnbmnbZ4JCUxDlliq38YX/UDIs1I8ho9CdqFdC1ZbnTSvwJgmiAAC4SbyTAgAA9IhrjaZTnqZTKy/fWvJbmpwXfGUbgqk355b0nXcWVPTb+0klXdsURGVS7VVTk0lXMWfnhVKRk5TvHJCfONC2zwT5pkqo5UAqsfDvsmGxfoxqENVYCSV7q9xiqNBJK3TSNCwHAGAVhE0AAABbLO5a7RmNac/oyqFFvhw0V0flm8Op164WlCv48sP2UGo07jSHUE3L9yqPJxKunB3S5DxyUvKdg/ITB7XUtCOSCfNySrP1q+ZVe0V582/JhkvSrDTV+BTjVYOnkVoAFTojCt100+PISSt0UpJxtvjVAgDQX4RNAAAAAyrlOUqNO7plvHs/qSiKNF8KlWtsbt6yfO+ta0uaK/pqzaSMpImE09zcvEOl1Fh8iJucG6PIGZGfHJGfvLV5XxTJBIvKpCPdyL4tGyzKBgu1P8ZflPWvy116p/JY7csjIxlFNtkQRI20BVXR8n03rcjEpWH9XgMAdgzCJgAAgG3MGKOxuKOxuKPbJruPC8JI15eWK6Xag6nZvK/XZ4u6vhS0Pde10mSitY9UQzBVvR3xhqzJuTGK3LSUnlapOLry2CiSCYuywXxzKOUvygQLtW3u0uXK9oale02HMW7HQKo1lKrsG6G/FABgIPHuBAAAsAM41tSCIinRdVw5qDY5b7jSXmOl1KUbJb3ybl6L5fYqnphj2vtIparBVHX7ZNJVwh3CflLGKHKSCpyk2uO6DiK/GkAtyvrVSqmGUMpW77ul92T9BZkuRw1tsq1iKnLSCt0Rhc5ocwWVTVA1BQDYEoRNAAAAqPEco5kRTzMjnqRk13HF5Sbn1SvvzTWGUwVf53NFZfO+SkF7P6kRz7ZdcS/TuJSvekU+zxniYMS4Ct1xhe641H2VZEUUyYRL9eV7Lcv5lgMrt3SlWjWV73wYObWKqNqyvS5L+0JnRLIrN7oHAKAbwiYAAACsW8K12jsa094VmpxHUaR8OWy72l5jtdT33strrujLby+U0ljcqYVQk22NzivL+MbjzvA3OTdGkZNQ4CQUaHr18VEgG+QbqqUWGiqo6kGVW7haGRP5HQ8T2nj7Er6WpX2VCqq0IpuUzBBWrAEANoSwCQAAAJvCGKORmKORmKMDKzQ5D6NI80tBUwiVbVm+d2FuSdc7NDm3RppINPeOyiRb+0u5Go07w9VPaiXGUeiOKnRX6TMlVaqmolJztZTfXjnllLPyim/JBIsyaq9Wi2QVOqlqMNUSSrmNQdVotWpq5SsxAgC2N8ImAAAA9JU1RuMJV+MJV7ev0uR8uZ9U6xX3snlf7y2U9drVguY7Njk3LX2kOizhS7lKukPW5Hw1xigycUU2rtCbWn18FMpUq6Za+0vVw6oFeeVcpaoqKnU8TGhiDdVSjYFU+3K+yBmhagoAthnCJgAAAGwLjjXKpDxlUp6U6T6uFFT7SbWGUtX7b11b0suXF5Xv0OQ84TaGUpX+UfszRalc1EjMasRzmm5TnjPcvaVaGavITStw02trhB6WOodSjUv6/Gtyly7JBosyav+ZRDKKqlVT3a7UFzojitzqkj8ToxE6APQZYRMAAACGSsyx2p2OaXd65aVa+XKguUKgXKHcFEpl85WG56/PFpQr+CoFuVW+XnW5oGc1EnOUroZRqZitbWsKqhrGjsSsYs4QV+3YmEIbU+itULK2LAplwkJDOFW/Ul9jY3R36e3KvrDY+TDG7RpIVZqitzRCN06PXzQAgLAJAAAAO1LKc5TyHO0fW7nJeWpsUm+9e1WLpUD5cqjFUqjFclC5LQVaLDffXi8Geme+pHwp1EIpUIcL8jVxrekSRnUJqBoDrJijuGOGY+mfsYqcEQXOiALtWn185LcFUjZYbG6KHszLXbpcrZrqXIsV2mRbILXcX6q1MXpkE1RNAcAaEDYBAAAAXRhjNBJ3NTPiaWbEW/fzoyhSKYi0UA2j8q0BVWNw1RBgXV0sK18dV1olrbJG7SFUbblf87ZUzCrdGGDFrBKuld2OAYpxFbrjCt3x1cdGkUxYbFrKZ6phVOOV+tzSe7KFN2TDQufDyGkIpBpCqLam6MtVU/xzC8DOxG8/AAAAYJMYYxR3jeKuXanN1IpKQSWkWqiGUfnGoKpDZdViKdRcYakSbJUDFf2VwyojVZf8rV5ZlarephsCrKRn5dgBD6uMUeQkFThJBZpZfXwUNPWaMg2BVOOV+tzSlcr+yO94mNAmmkKoqEMgVes5ZZM0QgcwNAibAAAAgAEWc6xiSauJ5MY+uvthVKuSqldQtVZVNQdY7y7UK6s6NVJvlXTtykFVUx+ryuN0Q4DlDlpYZRyF7phCd2z1sVEkE5UqoZPf4ep8y72myrOyxYsyQV5G7QFgJNtQKTXSsqyv/Up9suuvtAOArULYBAAAAAwx1xqNJVyNJTb2/CCMVCh3Xu7XXFVVvz+b9/VWuVQLsFZpW6X4cpP1Dlf8a1wimPJsU4CVrgZYfW2ybowiE1dg45K3hvq1KJAJ8p2v0hcsyPqV+145W6moikodDxOaWH35njva5Up91aboNkXVFIAtRdgEAAAAoCvHGqXjjtLxjV21LYwiFf2wPaBqCK7y5Uoz9eVt16pN1pefs1qTdW+5yXpLMJVuCbBSLc3Vl8duaZN14yhyRxW4o11alrcIS7Vm57VQym8OqpxyTm7xrWoj9E5VU6ZaLdUpkEordJuX+kUmRiN0ADeFsAkAAADAprHG1K78t9Em60tB1L70ryG4Wr5K4HIj9oVSoCuLZS2WAi2UQvnhymmVs9xkvSmYal7u11ZZ1RBcJV27eWGVjSm0MYXe5Opjo1AmLHQIpJobo7tLb1e2hUudD2O89kCq6Up9y8FUWqGTkszGgkgAw4uwCQAAAMDAMsYo4RolXKtMamPHKAWrV1Y1blsohcoVlmrPWVrDFQFrVVMNjdQbl/u1VVZ5m9Bk3VhFzogCZ0RBbA3jw3KX5XyLMsFCpSm6f13u0juVoEqd+3eFNqXQScu8O6qJIFJkXMm4ioyryDiS8RQZR5HxKpVd1cfLY+pjG+7bTtsbn0PABQwywiYAAAAAQ225yfrkBpusl4NI+daeVQ0VVq1h1WIp1LsLZS2WiloshSr4qzdZT3m2pYKqeblfuq2yajmsuokm69ZTaCcUehOrj40imbDYUi1VqaAyy8v5bCD5BdkwX7lCX+Q33VbuBx2X+q1XJNM5pKoGUW3Blu02tj3Qanxu5bbD2OrxJIclh0AHhE0AAAAAsALPMRp3XI3fRJP1fMOSv4WWCqu2IKsUaDZf1g+uVbbl19BkPeGatiWArcv9OjVdX756oLdak3VjFDlJBU5SgWY6Dpmenta12dmVjxNFkgKZKJCiskwUyITlSgjVFkytdD/oHmiF1fvhkky0WB27/DXqX6tbpdZ6tVdwrRZobTwU63ysSsUXTeAxSAibAAAAAGATOdZoNO5o9CaarBeqfamWg6mFTpVVDQHWtWKgt2+UattWaVulmGM6V1Z1DajqSwPTMUcxZ43VPcZIWq4KivegxukmREE9eFpDuKXIlwlXG9scbC0/z4bFzs8LfZm1tYpf/eXItgdTDRVYkfXWFIqtFo7VljN2XOroSrJUe4GwCQAAAAAGmTWmGu44kjbWZL3oR7UqqcWGRurdKqsWSoHeWyhXtwVabSWga6Wkd17WSK4xcqyRa41cq4b71e1G9fuN223D9uoxPGvkVI/htT6nepzGYzm2YZsxcp3lYzWPrRzLyjFWsrE+h15hNZxqqPZqrMBaUxDWUCWmoGsoZsOSFM23VXvVwrRNWeK4jj5dHUOx+nLGtVWMNVZ7EXr1C2ETAAAAAAwxY4ySnlHSs9IGmqxHUaRSEGmxHCrfUEG10FJZJS+uhcWCgiiSH1b+BGEkP1T1tvKnUNvfvL1+X7VjrFaR1QudQqu2kMysFJ4ZuY4axnQ4TjX4Wh7TOfwycoyVZx05Nt4WxDm2MTyrH79nV0KMIklh1wqs9gqu7kFY53CsMRRbklG+LRSrje3VEkc5XSuw1t2nq0Pz+k5BWqfn7cQljoRNAAAAAICujDGKu0Zx12pqhSbr09PTml2tZ9M6hVE9sKoFUlEkP6jcBtXtzWFVdXt1XHP41TK+wzHK1dsgVOV+1Ly96EcKwrD+nNp82sOzVS5k2BO2U1hmOoRiLZVj9fCrudKsFn6Z5YoyR451G46lthCuNTxz7fLx1RSQ1UK1hvG2U1gWhQ29uTpXeXWt9gpbn7dylVj7EsfGCjO/Jz+j+hLHSjhlrt0lTfz3nhx7UBE2AQAAAAAGkjVG1jHyNtbuqu+iqLlSq62Cq7WyK2re3hR+RZUrI3aqHGs8brBK5VjJj1SIwu7HaArpNv97ZI1Wrhxb3t5UOebKrQZgawnPOh+j+Wt6nY5lIrk2lGcCeSaQa0J5xpdrAjny5ZpQjrosSVwhFIunDmz+N7bPCJsAAAAAANgExlSCMk/bs3dQFFVCrsbQarlCLIjqQdhKlWOV4EotQVvL2KhDRVmXyrFyGKngh+3LLmvVbs3h2WYXlxktB2SeXOutGn651uhHb5nS/7hzkyfWZ4RNAAAAAACgjTH1hu7xfk9mg5orvjqFXp0Dq3LYuXKs8Vgr9hxrCr+ax5eCLSgZ67NtGTa9/PLL+ou/+AuFYagHH3xQJ0+e7PeUNp3v+8rn8yoUCpK6N4Fba3O49T5/LcfdyHNv9nUAAAAAANCNU60qGiSb0d9s0Gy7sCkMQ/35n/+5fud3fkeZTEa/9Vu/pePHj+uWW27p99Q21cWLF/WNb3yj39MYKJsRjG3WcYc9HLyZ427Gz8t1XQVBsOpx12K9wedmj+/FcbbDHLdifL++7iDNMxaLqVQqberX3UybPReO31vb+fUmEgktLS1t2vGBQTtfNyqRSKhYLPZ7GpKG53vaC3wvBs/Bgwd1+PDhfk9jU227sOn8+fPas2ePdu/eLUm6//77debMmaEPm8zFWe2N3SGta8Vp1HZvpV8za1vNGnWZQvvXWtOx1rx1pT3rW4W7oVW7UefX3bYp6rpnleeudU7dv2emw751v9JopVk1PDJdB3XcsdI81vK3oNujbsf1td6rRqzz72KnHRt4D9/UFeRdDr35q9bXe1736qtu9esyXbavtLUXX7eXozf2jEGx7p/5wL3Uzf7ZbrbBmtFO+ruP4TN8fxuH7xXh5m3+Z7Xt5d3vv03YNGhyuZwymUztcSaT0blz5/o4o60xuW+fDl+buKlqjajtX8MbOeFXjqtu/hhbpye/7gbkd2b7z7Z579qs9eey0vE28rPt9o/4jTPWKurLOugB+Qux6bb+dQ7Gb42tsLXfW8daBeHw9wwAbpZjrcIeVcz23wC8V/VsCgPwWnrmJl/LgHwrHGsVhn04V3ZM5c4GX+e2+/b0YcJb/CWnJ7ddFLNu2+4VRlH7b9JOZYHPPvusnn32WUnSE088oenp6U2f22aa/ui07vpvrnx/vRUbwM7jupwrwFpwrgBrw7kCrA3nCrA2O+Fc2XZhUyaTUTabrT3OZrOanJxsG3fixAmdOHGi9ngYmm/thCZiQC9wrgBrw7kCrA3nCrA2nCvA2gzTubJv376O2+0Wz+OmHT58WJcvX9aVK1fk+75efPFFHT9+vN/TAgAAAAAAgLZhZZPjOPqlX/olfeELX1AYhvqJn/gJHThwoN/TAgAAAAAAgLZh2CRJ9957r+69995+TwMAAAAAAAAttt0yOgAAAAAAAAwuwiYAAAAAAAD0DGETAAAAAAAAeoawCQAAAAAAAD1D2AQAAAAAAICeIWwCAAAAAABAzxA2AQAAAAAAoGcImwAAAAAAANAzhE0AAAAAAADoGRNFUdTvSQAAAAAAAGA4UNm0jXzuc5/r9xSAbYFzBVgbzhVgbThXgLXhXAHWZiecK4RNAAAAAAAA6BnCJgAAAAAAAPQMYdM2cuLEiX5PAdgWOFeAteFcAdaGcwVYG84VYG12wrlCg3AAAAAAAAD0DJVNAAAAAAAA6Bm33xNAZ88884xeeukljY+P68knn5QkLSws6KmnntLVq1c1MzOjz372s0qn032eKdBfs7Ozevrpp3Xt2jUZY3TixAn99E//tP7mb/5G//iP/6ixsTFJ0s///M/r3nvv7fNsgf569NFHlUgkZK2V4zh64okneG8BtP7PXV//+td1+vRpWWv1yCOP6NixY32cPbA1NvKZi3MFO9V6P3MN47lC2DSgPvaxj+mhhx7S008/Xdt26tQp3X333Tp58qROnTqlU6dO6eGHH+7jLIH+cxxHv/ALv6BDhw6pUCjoc5/7nD74wQ9Kkj7xiU/oZ37mZ/o8Q2Cw/P7v/37tHwQS7y2AtL7PXZcuXdKLL76oL37xi5qbm9Pjjz+uL3/5y7KWBQMYbuv9zMW5gp1urZ+5hvVc2d6zH2J33XVX2/8snzlzRg888IAk6YEHHtCZM2f6MTVgoExOTurQoUOSpGQyqf379yuXy/V5VsD2wXsLsL7PXWfOnNH9998vz/O0a9cu7dmzR+fPn9/yOQNbbb2fuThXgGY77X2FyqZt5Pr165qcnJRU+WV/48aNPs8IGCxXrlzRm2++qSNHjujs2bP6h3/4Bz3//PM6dOiQfvEXf5GlQYCkL3zhC5Kkj3/84zpx4gTvLUAX3c6NXC6no0eP1sZNTU3xnxzYcdbymYtzBTvdWj9zDeu5QtgEYCgUi0U9+eST+vSnP61UKqWf/Mmf1M/93M9Jkv76r/9af/mXf6nf+I3f6PMsgf56/PHHNTU1pevXr+sP/uAPtG/fvn5PCdh2uJAzdrq1fubiXMFOtp7PXMN6rrCMbhsZHx/X3NycJGlubq5p/Sewk/m+ryeffFI//uM/rg996EOSpImJCVlrZa3Vgw8+qDfeeKPPswT6b2pqSlLl/eS+++7T+fPneW8Buuh2bmQyGWWz2dq4XC5XO7eAYbeez1ycK9jJ1vOZa1jPFcKmbeT48eN67rnnJEnPPfec7rvvvj7PCOi/KIr0p3/6p9q/f78++clP1rYv/yKXpG9/+9s6cOBAP6YHDIxisahCoVC7/8orr+jgwYO8twBddDs3jh8/rhdffFHlcllXrlzR5cuXdeTIkX5OFdgS6/3MxbmCnWq9n7mG9Vwx0bDWbG1zX/rSl/Tqq69qfn5e4+Pj+tSnPqX77rtPTz31lGZnZzU9Pa3HHnuMHjTY8c6ePavf+73f08GDB2WMkVS55O4LL7ygixcvyhijmZkZ/eqv/mptjTSwE7333nv64z/+Y0lSEAT6yEc+op/92Z/V/Pw87y3Y8db7uetrX/uavvWtb8laq09/+tO65557+vwKgM23kc9cnCvYiTbymWsYzxXCJgAAAAAAAPQMy+gAAAAAAADQM4RNAAAAAAAA6BnCJgAAAAAAAPQMYRMAAAAAAAB6hrAJAAAAAAAAPUPYBAAA0GePPvqoXnnllX5PAwAAoCcImwAAAAAAANAzhE0AAAAAAADoGcImAACAAfL222/r0Ucf1QsvvNDvqQAAAGyI2+8JAAAAoOLChQv6oz/6I33mM5/Rj/3Yj/V7OgAAABtC2AQAADAAzp49q9OnT+s3f/M39YEPfKDf0wEAANgwltEBAAAMgG9+85u64447CJoAAMC2R9gEAAAwAH7lV35F2WxWX/nKV/o9FQAAgJtC2AQAADAAEomEfvu3f1uvvfaa/uqv/qrf0wEAANgwwiYAAIABMTIyot/93d/Vyy+/rK9+9av9ng4AAMCGmCiKon5PAgAAAAAAAMOByiYAAAAAAAD0DGETAAAAAAAAeoawCQAAAAAAAD1D2AQAAAAAAICeIWwCAAAAAABAzxA2AQAAAAAAoGcImwAAAAAAANAzhE0AAAAAAADoGcImAAAAAAAA9Mz/BxTkj74ZNaFNAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1440x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(20,6))\n",
    "k_str = [str(k) for k in k_list]\n",
    "plt.plot(k_str, run_k_anom_micro, label='k-Anonymity microaggregation')\n",
    "plt.plot(k_str, run_k_anom_mdav, label='k-Anonymity MDAV')\n",
    "plt.plot(k_str, run_t_clos, label='k-t-Closeness')\n",
    "plt.plot(k_str, run_dp_micro, label='Differential privacy microaggregation')\n",
    "plt.plot(k_str, run_dp_mdav, label='Differential privacy MDAV')\n",
    "plt.legend(loc='best', shadow=True, fontsize='12')\n",
    "plt.xlabel('k')\n",
    "plt.ylabel('Seconds')\n",
    "plt.title('Runtime')\n",
    "plt.xticks(k_str)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "venv",
   "language": "python",
   "name": "venv"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}