13_anomlay_detection.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "b66f8028",
   "metadata": {},
   "source": [
    "## Anomaly Detection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "a52345b0-e4cf-4448-a962-e93f609d4735",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: pyod in /opt/conda/lib/python3.8/site-packages (0.9.7)\n",
      "Requirement already satisfied: scikit-learn>=0.20.0 in /opt/conda/lib/python3.8/site-packages (from pyod) (1.1.2)\n",
      "Requirement already satisfied: numpy>=1.13 in /opt/conda/lib/python3.8/site-packages (from pyod) (1.19.5)\n",
      "Requirement already satisfied: scipy>=1.3.1 in /opt/conda/lib/python3.8/site-packages (from pyod) (1.7.3)\n",
      "Requirement already satisfied: matplotlib in /opt/conda/lib/python3.8/site-packages (from pyod) (3.5.1)\n",
      "Requirement already satisfied: joblib in /opt/conda/lib/python3.8/site-packages (from pyod) (1.0.1)\n",
      "Requirement already satisfied: numba>=0.35 in /opt/conda/lib/python3.8/site-packages (from pyod) (0.54.1)\n",
      "Requirement already satisfied: six in /opt/conda/lib/python3.8/site-packages (from pyod) (1.15.0)\n",
      "Requirement already satisfied: statsmodels in /opt/conda/lib/python3.8/site-packages (from pyod) (0.13.1)\n",
      "Requirement already satisfied: llvmlite<0.38,>=0.37.0rc1 in /opt/conda/lib/python3.8/site-packages (from numba>=0.35->pyod) (0.37.0)\n",
      "Requirement already satisfied: setuptools in /opt/conda/lib/python3.8/site-packages (from numba>=0.35->pyod) (60.2.0)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in /opt/conda/lib/python3.8/site-packages (from scikit-learn>=0.20.0->pyod) (3.0.0)\n",
      "Requirement already satisfied: fonttools>=4.22.0 in /opt/conda/lib/python3.8/site-packages (from matplotlib->pyod) (4.28.5)\n",
      "Requirement already satisfied: cycler>=0.10 in /opt/conda/lib/python3.8/site-packages (from matplotlib->pyod) (0.11.0)\n",
      "Requirement already satisfied: pillow>=6.2.0 in /opt/conda/lib/python3.8/site-packages (from matplotlib->pyod) (9.0.0)\n",
      "Requirement already satisfied: packaging>=20.0 in /opt/conda/lib/python3.8/site-packages (from matplotlib->pyod) (21.3)\n",
      "Requirement already satisfied: python-dateutil>=2.7 in /opt/conda/lib/python3.8/site-packages (from matplotlib->pyod) (2.8.2)\n",
      "Requirement already satisfied: pyparsing>=2.2.1 in /opt/conda/lib/python3.8/site-packages (from matplotlib->pyod) (3.0.6)\n",
      "Requirement already satisfied: kiwisolver>=1.0.1 in /opt/conda/lib/python3.8/site-packages (from matplotlib->pyod) (1.3.2)\n",
      "Requirement already satisfied: pandas>=0.25 in /opt/conda/lib/python3.8/site-packages (from statsmodels->pyod) (1.3.5)\n",
      "Requirement already satisfied: patsy>=0.5.2 in /opt/conda/lib/python3.8/site-packages (from statsmodels->pyod) (0.5.2)\n",
      "Requirement already satisfied: pytz>=2017.3 in /opt/conda/lib/python3.8/site-packages (from pandas>=0.25->statsmodels->pyod) (2021.3)\n",
      "\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\n",
      "\u001b[33mWARNING: You are using pip version 21.3.1; however, version 22.2.2 is available.\n",
      "You should consider upgrading via the '/opt/conda/bin/python3.8 -m pip install --upgrade pip' command.\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "!pip install pyod"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c1b6b410-62ed-4280-b846-dabc8c30333f",
   "metadata": {},
   "source": [
    "## 가상데이터 생성"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "679f74f0-e954-4439-9754-ffe3a92c5fdc",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.8/site-packages/pyod/utils/data.py:186: FutureWarning: behaviour=\"old\" is deprecated and will be removed in version 0.9.0. Please use behaviour=\"new\", which makes the returned datasets in the order of X_train, X_test, y_train, y_test.\n",
      "  warn('behaviour=\"old\" is deprecated and will be removed '\n"
     ]
    }
   ],
   "source": [
    "from pyod.utils.data import generate_data\n",
    "from pyod.utils.example import visualize\n",
    "contamination = 0.05  # percentage of outliers\n",
    "n_train = 2000  # number of training points\n",
    "n_test = 1000  # number of testing points\n",
    "\n",
    "X_train, y_train, X_test, y_test = generate_data(\n",
    "    n_train=n_train, n_test=n_test, contamination=contamination)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "8148f098-3716-4192-9c97-5b9fb8dea3d0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f7f1b2bad90>"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.scatter(X_train[:,0],X_train[:,1],c=y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "25da7235-fb52-4778-955d-000107750f69",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f7f131566a0>"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.scatter(X_test[:,0],X_test[:,1],c=y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e14f0422-fe4f-4243-89ca-8761711a779d",
   "metadata": {},
   "source": [
    "## 이상탐지 모듈 불러오기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "a4e75ab8-2d9e-4b5f-95ed-d97663ad1cd6",
   "metadata": {},
   "outputs": [],
   "source": [
    "from pyod.models.iforest import IForest\n",
    "from pyod.models.knn import KNN\n",
    "from pyod.models.pca import PCA\n",
    "from pyod.models.lof import LOF"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8ed48816-9fda-4a09-a234-158fe6f34a3d",
   "metadata": {},
   "source": [
    "## Fully Unsupervised Approach"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "037497ba-36ac-4558-b21f-fd20f14d2736",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNN(algorithm='auto', contamination=0.1, leaf_size=30, method='largest',\n",
       "  metric='minkowski', metric_params=None, n_jobs=1, n_neighbors=5, p=2,\n",
       "  radius=1.0)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = KNN(n_neighbors=5)\n",
    "# Suppose that we don't know the true label\n",
    "# In this case, we just learn from whole data\n",
    "clf.fit(X_train) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "85c46a4b-68f3-4154-b1cc-e444954e3c11",
   "metadata": {},
   "outputs": [],
   "source": [
    "# making a prediction\n",
    "y_pred=clf.predict(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6e933664-c04e-4485-9889-e93850bdab51",
   "metadata": {},
   "source": [
    "## Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "d23faf07-6bdb-4d3c-bd98-5e9f7f606346",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy_score: 0.945\n",
      "precision_score: 0.47619047619047616\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9995789473684211\n",
      "Prec@N: 0.94\n"
     ]
    }
   ],
   "source": [
    "## Since we know the true label\n",
    "## Let's evaluate\n",
    "from sklearn.metrics import accuracy_score, precision_score, recall_score, roc_auc_score\n",
    "from pyod.utils.utility import precision_n_scores\n",
    "\n",
    "# accuracy\n",
    "print(f\"accuracy_score: {accuracy_score(y_test,y_pred)}\") #(tp+tn)/(tp+tn+fp+fn)\n",
    "print(f\"precision_score: {precision_score(y_test,y_pred)}\") #(tp)/(tp+fp)\n",
    "print(f\"recall_score: {recall_score(y_test,y_pred)}\") #(tp)/(tp+fn)\n",
    "\n",
    "score = clf.decision_function(X_test)\n",
    "print(f\"ROC_AUC: {roc_auc_score(y_test,score)}\") #(tp)/(tp+fn)\n",
    "print(f\"Prec@N: {precision_n_scores(y_test,score)}\")\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "f717daeb-fa47-40aa-ba68-8bdf1d98cd1a",
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate(y_test, y_pred, score):    \n",
    "    print(f\"accuracy_score: {accuracy_score(y_test,y_pred)}\") #(tp+tn)/(tp+tn+fp+fn)\n",
    "    print(f\"precision_score: {precision_score(y_test,y_pred)}\") #(tp)/(tp+fp)\n",
    "    print(f\"recall_score: {recall_score(y_test,y_pred)}\") #(tp)/(tp+fn)\n",
    "    print(f\"ROC_AUC: {roc_auc_score(y_test,score)}\") #(tp)/(tp+fn)\n",
    "    print(f\"Prec@N: {precision_n_scores(y_test,score)}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bcc71607-2cec-484d-a968-0b237f23a718",
   "metadata": {},
   "source": [
    "## Semi-supervised"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "e71ba6ea-a1a9-4ad4-a162-1537afedfad4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNN(algorithm='auto', contamination=0.1, leaf_size=30, method='largest',\n",
       "  metric='minkowski', metric_params=None, n_jobs=1, n_neighbors=5, p=2,\n",
       "  radius=1.0)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = KNN(n_neighbors=5)\n",
    "clf.fit(X_train[y_train==0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "bf01b8d3-a20b-4cfa-992d-910c0f9c1dda",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = clf.predict(X_test)\n",
    "score = clf.decision_function(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "0522053e-8a49-4c21-b7da-b5003ea07abe",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy_score: 0.891\n",
      "precision_score: 0.31446540880503143\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9995789473684211\n",
      "Prec@N: 0.94\n"
     ]
    }
   ],
   "source": [
    "evaluate(y_test,y_pred,score)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "273864b8-c714-4a78-bbeb-77f075a22129",
   "metadata": {},
   "source": [
    "## ParamTuning Example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "b1f61515-378a-4863-b06b-aaa37db62d06",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "==========\n",
      "contamination: 0.01 and n_nighbors: 1\n",
      "accuracy_score: 0.987\n",
      "precision_score: 0.8032786885245902\n",
      "recall_score: 0.98\n",
      "ROC_AUC: 0.9990947368421053\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.01 and n_nighbors: 3\n",
      "accuracy_score: 0.984\n",
      "precision_score: 0.7575757575757576\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9994526315789474\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.01 and n_nighbors: 5\n",
      "accuracy_score: 0.986\n",
      "precision_score: 0.78125\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9995789473684211\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.01 and n_nighbors: 7\n",
      "accuracy_score: 0.988\n",
      "precision_score: 0.8166666666666667\n",
      "recall_score: 0.98\n",
      "ROC_AUC: 0.9994736842105263\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.01 and n_nighbors: 9\n",
      "accuracy_score: 0.985\n",
      "precision_score: 0.7777777777777778\n",
      "recall_score: 0.98\n",
      "ROC_AUC: 0.9994315789473683\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.05 and n_nighbors: 1\n",
      "accuracy_score: 0.944\n",
      "precision_score: 0.4716981132075472\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9990947368421053\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.05 and n_nighbors: 3\n",
      "accuracy_score: 0.947\n",
      "precision_score: 0.4854368932038835\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9994526315789474\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.05 and n_nighbors: 5\n",
      "accuracy_score: 0.947\n",
      "precision_score: 0.4854368932038835\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9995789473684211\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.05 and n_nighbors: 7\n",
      "accuracy_score: 0.95\n",
      "precision_score: 0.5\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9994736842105263\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.05 and n_nighbors: 9\n",
      "accuracy_score: 0.951\n",
      "precision_score: 0.5050505050505051\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9994315789473683\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.1 and n_nighbors: 1\n",
      "accuracy_score: 0.901\n",
      "precision_score: 0.33557046979865773\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9990947368421053\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.1 and n_nighbors: 3\n",
      "accuracy_score: 0.902\n",
      "precision_score: 0.33783783783783783\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9994526315789474\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.1 and n_nighbors: 5\n",
      "accuracy_score: 0.891\n",
      "precision_score: 0.31446540880503143\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9995789473684211\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.1 and n_nighbors: 7\n",
      "accuracy_score: 0.89\n",
      "precision_score: 0.3125\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9994736842105263\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.1 and n_nighbors: 9\n",
      "accuracy_score: 0.892\n",
      "precision_score: 0.31645569620253167\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9994315789473683\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.15 and n_nighbors: 1\n",
      "accuracy_score: 0.848\n",
      "precision_score: 0.24752475247524752\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9990947368421053\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.15 and n_nighbors: 3\n",
      "accuracy_score: 0.865\n",
      "precision_score: 0.2702702702702703\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9994526315789474\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.15 and n_nighbors: 5\n",
      "accuracy_score: 0.858\n",
      "precision_score: 0.2604166666666667\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9995789473684211\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.15 and n_nighbors: 7\n",
      "accuracy_score: 0.853\n",
      "precision_score: 0.25380710659898476\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9994736842105263\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.15 and n_nighbors: 9\n",
      "accuracy_score: 0.856\n",
      "precision_score: 0.25773195876288657\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9994315789473683\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.2 and n_nighbors: 1\n",
      "accuracy_score: 0.813\n",
      "precision_score: 0.2109704641350211\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9990947368421053\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.2 and n_nighbors: 3\n",
      "accuracy_score: 0.816\n",
      "precision_score: 0.21367521367521367\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9994526315789474\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.2 and n_nighbors: 5\n",
      "accuracy_score: 0.818\n",
      "precision_score: 0.21551724137931033\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9995789473684211\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.2 and n_nighbors: 7\n",
      "accuracy_score: 0.816\n",
      "precision_score: 0.21367521367521367\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9994736842105263\n",
      "Prec@N: 0.94\n",
      "==========\n",
      "contamination: 0.2 and n_nighbors: 9\n",
      "accuracy_score: 0.82\n",
      "precision_score: 0.21739130434782608\n",
      "recall_score: 1.0\n",
      "ROC_AUC: 0.9994315789473683\n",
      "Prec@N: 0.94\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_score\n",
    "KNN()\n",
    "contamination=[0.01,0.05,0.1,0.15,0.2] #hyper parameters for contamination\n",
    "n_neighbors=[1,3,5,7,9] # hyper parameters for neighbors\n",
    "for c in contamination:\n",
    "    for n in n_neighbors:\n",
    "        print(\"=\"*10)\n",
    "        clf = KNN(n_neighbors=n,contamination=c)\n",
    "        clf.fit(X_train[y_train==0])\n",
    "        y_pred = clf.predict(X_test)\n",
    "        score = clf.decision_function(X_test)\n",
    "        print(f\"contamination: {c} and n_nighbors: {n}\")\n",
    "        evaluate(y_test,y_pred,score)\n",
    "\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "70d37bc5-a1c1-44cb-971a-54653fbdfd5e",
   "metadata": {},
   "source": [
    "## For Actual Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "8640cc19-366b-4499-bd9c-632ccb5a8d85",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-09-28T06:04:05.270395Z",
     "start_time": "2022-09-28T06:04:04.786223Z"
    }
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "data = np.load('13_fraud.npz', allow_pickle=True)\n",
    "# data = np.load('34_smtp.npz', allow_pickle=True)\n",
    "# data = np.load('16_http.npz', allow_pickle=True)\n",
    "# df=pd.read_csv('f.csv')\n",
    "# df\n",
    "X, y = data['X'], data['y']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "0ceea63e",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-09-28T06:06:18.488509Z",
     "start_time": "2022-09-28T06:06:18.441506Z"
    }
   },
   "outputs": [],
   "source": [
    "anomal_X, anomal_y = X[y==1], y[y==1]\n",
    "normal_X, normal_y = X[y==0], y[y==0]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "c7362f94",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-09-28T06:06:26.041512Z",
     "start_time": "2022-09-28T06:06:25.996238Z"
    }
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(normal_X, normal_y, shuffle=True, test_size=0.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "2709c82d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-09-28T06:08:30.358754Z",
     "start_time": "2022-09-28T06:08:30.352063Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(199020, 29)"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "6373352d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-09-28T06:10:03.316348Z",
     "start_time": "2022-09-28T06:10:03.304949Z"
    }
   },
   "outputs": [],
   "source": [
    "X_test=np.vstack([X_test,anomal_X])\n",
    "y_test=np.hstack([y_test,anomal_y])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "4fb3b8da",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-09-28T06:11:05.133967Z",
     "start_time": "2022-09-28T06:11:05.126329Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(88739, 29)"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "09e02958",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-09-27T12:20:16.464915Z",
     "start_time": "2022-09-27T12:20:16.455478Z"
    }
   },
   "source": [
    "## Using supervised model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "3467e1ab",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-09-28T06:10:11.411961Z",
     "start_time": "2022-09-28T06:10:11.106454Z"
    }
   },
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Found input variables with inconsistent numbers of samples: [85787, 88739]",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m/tmp/ipykernel_11646/2390960493.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0my_pred\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mclf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m \u001b[0mscore\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mclf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m \u001b[0mevaluate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0my_pred\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mscore\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/tmp/ipykernel_11646/3609646086.py\u001b[0m in \u001b[0;36mevaluate\u001b[0;34m(y_test, y_pred, score)\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mevaluate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_pred\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mscore\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf\"accuracy_score: {accuracy_score(y_test,y_pred)}\"\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m#(tp+tn)/(tp+tn+fp+fn)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf\"precision_score: {precision_score(y_test,y_pred)}\"\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m#(tp)/(tp+fp)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf\"recall_score: {recall_score(y_test,y_pred)}\"\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m#(tp)/(tp+fn)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf\"ROC_AUC: {roc_auc_score(y_test,score)}\"\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m#(tp)/(tp+fn)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.8/site-packages/sklearn/metrics/_classification.py\u001b[0m in \u001b[0;36maccuracy_score\u001b[0;34m(y_true, y_pred, normalize, sample_weight)\u001b[0m\n\u001b[1;32m    209\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    210\u001b[0m     \u001b[0;31m# Compute accuracy for each possible representation\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 211\u001b[0;31m     \u001b[0my_type\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_true\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_pred\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_check_targets\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_true\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_pred\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    212\u001b[0m     \u001b[0mcheck_consistent_length\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_true\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_pred\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msample_weight\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    213\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0my_type\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstartswith\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"multilabel\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.8/site-packages/sklearn/metrics/_classification.py\u001b[0m in \u001b[0;36m_check_targets\u001b[0;34m(y_true, y_pred)\u001b[0m\n\u001b[1;32m     82\u001b[0m     \u001b[0my_pred\u001b[0m \u001b[0;34m:\u001b[0m \u001b[0marray\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mindicator\u001b[0m \u001b[0mmatrix\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     83\u001b[0m     \"\"\"\n\u001b[0;32m---> 84\u001b[0;31m     \u001b[0mcheck_consistent_length\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_true\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_pred\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     85\u001b[0m     \u001b[0mtype_true\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtype_of_target\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_true\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput_name\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"y_true\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     86\u001b[0m     \u001b[0mtype_pred\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtype_of_target\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_pred\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput_name\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"y_pred\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.8/site-packages/sklearn/utils/validation.py\u001b[0m in \u001b[0;36mcheck_consistent_length\u001b[0;34m(*arrays)\u001b[0m\n\u001b[1;32m    385\u001b[0m     \u001b[0muniques\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0munique\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlengths\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    386\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0muniques\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 387\u001b[0;31m         raise ValueError(\n\u001b[0m\u001b[1;32m    388\u001b[0m             \u001b[0;34m\"Found input variables with inconsistent numbers of samples: %r\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    389\u001b[0m             \u001b[0;34m%\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ml\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mlengths\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: Found input variables with inconsistent numbers of samples: [85787, 88739]"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from lightgbm import LGBMClassifier\n",
    "from pyod.utils.utility import precision_n_scores\n",
    "clf = LGBMClassifier()\n",
    "clf.fit(X_train,y_train)\n",
    "y_pred=clf.predict(X_test)\n",
    "score = clf.predict_proba(X_test)[:,1]\n",
    "evaluate(y_test,y_pred,score)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "id": "8d452f1b",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-09-27T12:16:04.805496Z",
     "start_time": "2022-09-27T12:16:04.800025Z"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_auc_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "d0673efa",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-09-27T12:16:33.410109Z",
     "start_time": "2022-09-27T12:16:27.561257Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "==========\n",
      "contamination: 0.0005, n_estimators: 50\n",
      "accuracy_score: 0.997940147232658\n",
      "precision_score: 0.26666666666666666\n",
      "recall_score: 0.10810810810810811\n",
      "ROC_AUC: 0.9448416703238205\n",
      "Prec@N: 0.2635135135135135\n",
      "==========\n",
      "contamination: 0.0005, n_estimators: 100\n",
      "accuracy_score: 0.9978114064346991\n",
      "precision_score: 0.1320754716981132\n",
      "recall_score: 0.0472972972972973\n",
      "ROC_AUC: 0.9338393936465335\n",
      "Prec@N: 0.18243243243243243\n",
      "==========\n",
      "contamination: 0.0005, n_estimators: 150\n",
      "accuracy_score: 0.9979869620682794\n",
      "precision_score: 0.33783783783783783\n",
      "recall_score: 0.16891891891891891\n",
      "ROC_AUC: 0.9441870265834156\n",
      "Prec@N: 0.27702702702702703\n",
      "==========\n",
      "contamination: 0.0005, n_estimators: 200\n",
      "accuracy_score: 0.997905036105942\n",
      "precision_score: 0.26153846153846155\n",
      "recall_score: 0.11486486486486487\n",
      "ROC_AUC: 0.9421481567152474\n",
      "Prec@N: 0.28378378378378377\n",
      "==========\n",
      "contamination: 0.001, n_estimators: 50\n",
      "accuracy_score: 0.9978114064346991\n",
      "precision_score: 0.3488372093023256\n",
      "recall_score: 0.30405405405405406\n",
      "ROC_AUC: 0.9473089024894523\n",
      "Prec@N: 0.3108108108108108\n",
      "==========\n",
      "contamination: 0.001, n_estimators: 100\n",
      "accuracy_score: 0.9977645915990777\n",
      "precision_score: 0.32231404958677684\n",
      "recall_score: 0.2635135135135135\n",
      "ROC_AUC: 0.9506312353152732\n",
      "Prec@N: 0.28378378378378377\n",
      "==========\n",
      "contamination: 0.001, n_estimators: 150\n",
      "accuracy_score: 0.9977177767634563\n",
      "precision_score: 0.312\n",
      "recall_score: 0.2635135135135135\n",
      "ROC_AUC: 0.9393248075439293\n",
      "Prec@N: 0.2905405405405405\n",
      "==========\n",
      "contamination: 0.001, n_estimators: 200\n",
      "accuracy_score: 0.9975656285476867\n",
      "precision_score: 0.2413793103448276\n",
      "recall_score: 0.1891891891891892\n",
      "ROC_AUC: 0.9441731637258926\n",
      "Prec@N: 0.24324324324324326\n",
      "==========\n",
      "contamination: 0.01, n_estimators: 50\n",
      "accuracy_score: 0.9892208840981707\n",
      "precision_score: 0.07386990077177508\n",
      "recall_score: 0.4527027027027027\n",
      "ROC_AUC: 0.9401656492649516\n",
      "Prec@N: 0.16891891891891891\n",
      "==========\n",
      "contamination: 0.01, n_estimators: 100\n",
      "accuracy_score: 0.9893379211872242\n",
      "precision_score: 0.0866738894907909\n",
      "recall_score: 0.5405405405405406\n",
      "ROC_AUC: 0.9480154725333223\n",
      "Prec@N: 0.2635135135135135\n",
      "==========\n",
      "contamination: 0.01, n_estimators: 150\n",
      "accuracy_score: 0.9894198471495617\n",
      "precision_score: 0.08185840707964602\n",
      "recall_score: 0.5\n",
      "ROC_AUC: 0.9409640310337889\n",
      "Prec@N: 0.22297297297297297\n",
      "==========\n",
      "contamination: 0.01, n_estimators: 200\n",
      "accuracy_score: 0.9897124398721955\n",
      "precision_score: 0.1005464480874317\n",
      "recall_score: 0.6216216216216216\n",
      "ROC_AUC: 0.9471922564454367\n",
      "Prec@N: 0.28378378378378377\n",
      "==========\n",
      "contamination: 0.1, n_estimators: 50\n",
      "accuracy_score: 0.8996055850098896\n",
      "precision_score: 0.014959723820483314\n",
      "recall_score: 0.8783783783783784\n",
      "ROC_AUC: 0.9446719889477379\n",
      "Prec@N: 0.17567567567567569\n",
      "==========\n",
      "contamination: 0.1, n_estimators: 100\n",
      "accuracy_score: 0.8989150661844738\n",
      "precision_score: 0.014858841010401188\n",
      "recall_score: 0.8783783783783784\n",
      "ROC_AUC: 0.9402315176422685\n",
      "Prec@N: 0.25675675675675674\n",
      "==========\n",
      "contamination: 0.1, n_estimators: 150\n",
      "accuracy_score: 0.8993481034139719\n",
      "precision_score: 0.014921946740128558\n",
      "recall_score: 0.8783783783783784\n",
      "ROC_AUC: 0.9463470182181714\n",
      "Prec@N: 0.2905405405405405\n",
      "==========\n",
      "contamination: 0.1, n_estimators: 200\n",
      "accuracy_score: 0.900682326229182\n",
      "precision_score: 0.015119795301232844\n",
      "recall_score: 0.8783783783783784\n",
      "ROC_AUC: 0.943122596774628\n",
      "Prec@N: 0.24324324324324326\n"
     ]
    }
   ],
   "source": [
    "contamination=[0.0005,0.001,0.01,0.1]\n",
    "n_estimators=[50,100,150,200]\n",
    "for c in contamination:\n",
    "    for n in n_estimators:\n",
    "        clf = IForest(contamination=c,n_estimators=n)\n",
    "        clf.fit(X_train[y_train==0])\n",
    "        y_pred=clf.predict(X_test)\n",
    "        score=clf.predict_proba(X_test)[:,1]\n",
    "        print('='*10)\n",
    "        print(f'contamination: {c}, n_estimators: {n}')        \n",
    "        evaluate(y_test,y_pred,score)\n",
    "        "
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.10"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": false,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}