3.2.random_forestPruned .ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np \n",
    "import pandas as pd \n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import RandomForestClassifier as rfc\n",
    "from sklearn.metrics import precision_score, recall_score, roc_auc_score, roc_curve, auc\n",
    "from sklearn.metrics import accuracy_score\n",
    "import hyperopt as hp "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "size brca, ovca 1080 579\n",
      "normal size 1659\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:15: FutureWarning: Sorting because non-concatenation axis is not aligned. A future version\n",
      "of pandas will change to not sort by default.\n",
      "\n",
      "To accept the future behavior, pass 'sort=False'.\n",
      "\n",
      "To retain the current behavior and silence the warning, pass 'sort=True'.\n",
      "\n",
      "  from ipykernel import kernelapp as app\n"
     ]
    }
   ],
   "source": [
    "#data sets being combined \n",
    "def get_datasets(std_size=False):\n",
    "    # My python only likes absolute paths \n",
    "    brca = pd.read_csv('/Users/danielraad/Desktop/data/Gistic2_CopyNumber_BRCA', sep='\\t')\n",
    "    ovca = pd.read_csv('/Users/danielraad/Desktop/data/Gistic2_CopyNumber_OVCA', sep='\\t')\n",
    "    brca.set_index('Gene Symbol', inplace=True)\n",
    "    ovca.set_index('Gene Symbol', inplace=True)\n",
    "    brca = brca.transpose()\n",
    "    ovca = ovca.transpose()\n",
    "    brca['BRCA'] = 1\n",
    "    ovca['BRCA'] = 0\n",
    "    print('size brca, ovca', len(brca), len(ovca))\n",
    "    if not std_size:\n",
    "        print('normal size', len(brca)+len(ovca))\n",
    "        df = pd.concat([brca, ovca[1:]])\n",
    "    elif std_size:\n",
    "        print('sampled size', len(ovca)+len(ovca))\n",
    "        df = pd.concat([brca[0:len(ovca)], ovca[1:]])\n",
    "    return df\n",
    "df = get_datasets()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "#seperate the target variable from the other variables\n",
    "y = df[\"BRCA\"]\n",
    "X = df[df.columns[:-1]]\n",
    "X_train, X_test, y_train, y_tests = train_test_split(X, y, test_size = 0.33, random_state = 42)\n",
    "y_red = pd.Series(y)\n",
    "features = list(X_train.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Pruning data, finding the best values to use in the parameters \n",
    "\n",
    "#will do both manual technique and using hyperopt "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "#will see what is the best number of trees in the forest to use \n",
    "# value is between 20 and\n",
    "n_estimators = [1,2,4,8,16,32,64,100,200]\n",
    "\n",
    "train_results = [] \n",
    "test_results = [] \n",
    "for estimator in n_estimators: \n",
    "    rf = RandomForestClassifier(n_estimators=estimator, n_jobs=-1)\n",
    "    rf.fit(X_train, y_train)\n",
    "    \n",
    "    train_pred = rf.predict(X_train)\n",
    "    \n",
    "    false_positive_rate, true_positive_rate, thresholds = roc_curve(y_train, train_pred)\n",
    "    roc_auc = auc(false_positive_rate, true_positive_rate)\n",
    "    train_results.append(roc_auc)\n",
    "    \n",
    "    y_pred = rf.predict(X_test)\n",
    "    false_positive_rate, true_positive_rate, thresholds = roc_curve(y_tests, y_pred)\n",
    "    roc_auc = auc(false_positive_rate, true_positive_rate)\n",
    "    test_results.append(roc_auc)\n",
    "    \n",
    "from matplotlib.legend_handler import HandlerLine2D\n",
    "line1, = plt.plot(n_estimators, train_results, 'b', label=\"Train AUC\")\n",
    "line2, = plt.plot(n_estimators, test_results, 'r', label=\"Test AUC\")\n",
    "plt.legend(handler_map={line1: HandlerLine2D(numpoints=2)})\n",
    "plt.ylabel('AUC score')\n",
    "plt.xlabel('n_estimators')\n",
    "plt.show()\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n"
     ]
    },
    {
     "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": [
    "#finding the optimum depth \n",
    "max_depths = np.linspace(1, 32, 32, endpoint=True)\n",
    "train_results = []\n",
    "test_results = []\n",
    "for max_depth in max_depths:\n",
    "   rf = RandomForestClassifier(max_depth=max_depth, n_jobs=-1)\n",
    "   rf.fit(X_train, y_train)\n",
    "   train_pred = rf.predict(X_train)\n",
    "   false_positive_rate, true_positive_rate, thresholds = roc_curve(y_train, train_pred)\n",
    "   roc_auc = auc(false_positive_rate, true_positive_rate)\n",
    "   train_results.append(roc_auc)\n",
    "   y_pred = rf.predict(X_test)\n",
    "   false_positive_rate, true_positive_rate, thresholds = roc_curve(y_tests, y_pred)\n",
    "   roc_auc = auc(false_positive_rate, true_positive_rate)\n",
    "   test_results.append(roc_auc)\n",
    "from matplotlib.legend_handler import HandlerLine2D\n",
    "line1, = plt.plot(max_depths, train_results, 'b', label=\"Train AUC\")\n",
    "line2, = plt.plot(max_depths, test_results, 'r', label=\"Test AUC\")\n",
    "plt.legend(handler_map={line1: HandlerLine2D(numpoints=2)})\n",
    "plt.ylabel('AUC score')\n",
    "plt.xlabel('Tree depth')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "## tuning techniques from: \n",
    "## https://towardsdatascience.com/hyperparameter-tuning-the-random-forest-in-python-using-scikit-learn-28d2aa77dd74"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "rf = RandomForestClassifier(random_state = 42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Parameters currently in use:\n",
      "\n",
      "{'bootstrap': True, 'class_weight': None, 'criterion': 'gini', 'max_depth': None, 'max_features': 'auto', 'max_leaf_nodes': None, 'min_impurity_decrease': 0.0, 'min_impurity_split': None, 'min_samples_leaf': 1, 'min_samples_split': 2, 'min_weight_fraction_leaf': 0.0, 'n_estimators': 'warn', 'n_jobs': None, 'oob_score': False, 'random_state': 42, 'verbose': 0, 'warm_start': False}\n"
     ]
    }
   ],
   "source": [
    "print('Parameters currently in use:\\n')\n",
    "print(rf.get_params())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'n_estimators': [200, 400, 600, 800, 1000, 1200, 1400, 1600, 1800, 2000], 'max_features': ['auto', 'sqrt'], 'max_depth': [10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, None], 'min_samples_split': [2, 5, 10], 'min_samples_leaf': [1, 2, 4], 'bootstrap': [True, False]}\n"
     ]
    }
   ],
   "source": [
    "#this is going to random search all of these different paramters available to random forest and give me an idea \n",
    "#of where to look in my data\n",
    "from sklearn.model_selection import RandomizedSearchCV \n",
    "\n",
    "#number of trees in random forest \n",
    "n_estimators = [int(x) for x in np.linspace(start = 200, stop = 2000, num = 10) ]\n",
    "max_features = ['auto', 'sqrt']\n",
    "max_depth = [int(x) for x in np.linspace(10,110, num = 11)]\n",
    "max_depth.append(None)\n",
    "min_samples_split = [2,5,10]\n",
    "min_samples_leaf = [1,2,4]\n",
    "bootstrap = [True, False]\n",
    "\n",
    "random_grid = {'n_estimators': n_estimators,\n",
    "               'max_features': max_features,\n",
    "               'max_depth': max_depth,\n",
    "               'min_samples_split': min_samples_split,\n",
    "               'min_samples_leaf': min_samples_leaf,\n",
    "               'bootstrap': bootstrap}\n",
    "\n",
    "print(random_grid)\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 100 candidates, totalling 300 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  25 tasks      | elapsed:  5.5min\n",
      "[Parallel(n_jobs=-1)]: Done 146 tasks      | elapsed: 28.2min\n",
      "[Parallel(n_jobs=-1)]: Done 300 out of 300 | elapsed: 55.5min finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomizedSearchCV(cv=3, error_score='raise-deprecating',\n",
       "                   estimator=RandomForestClassifier(bootstrap=True,\n",
       "                                                    class_weight=None,\n",
       "                                                    criterion='gini',\n",
       "                                                    max_depth=None,\n",
       "                                                    max_features='auto',\n",
       "                                                    max_leaf_nodes=None,\n",
       "                                                    min_impurity_decrease=0.0,\n",
       "                                                    min_impurity_split=None,\n",
       "                                                    min_samples_leaf=1,\n",
       "                                                    min_samples_split=2,\n",
       "                                                    min_weight_fraction_leaf=0.0,\n",
       "                                                    n_estimators='warn',\n",
       "                                                    n_jobs=None,\n",
       "                                                    oob_sc...\n",
       "                   param_distributions={'bootstrap': [True, False],\n",
       "                                        'max_depth': [10, 20, 30, 40, 50, 60,\n",
       "                                                      70, 80, 90, 100, 110,\n",
       "                                                      None],\n",
       "                                        'max_features': ['auto', 'sqrt'],\n",
       "                                        'min_samples_leaf': [1, 2, 4],\n",
       "                                        'min_samples_split': [2, 5, 10],\n",
       "                                        'n_estimators': [200, 400, 600, 800,\n",
       "                                                         1000, 1200, 1400, 1600,\n",
       "                                                         1800, 2000]},\n",
       "                   pre_dispatch='2*n_jobs', random_state=42, refit=True,\n",
       "                   return_train_score=False, scoring=None, verbose=2)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_random = RandomizedSearchCV(estimator = rf, param_distributions = random_grid, n_iter = 100, cv = 3, verbose=2, random_state=42, n_jobs= -1)\n",
    "rf_random.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'n_estimators': 600, 'min_samples_split': 5, 'min_samples_leaf': 2, 'max_features': 'auto', 'max_depth': 70, 'bootstrap': False}\n"
     ]
    }
   ],
   "source": [
    "print(rf_random.best_params_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "#this provides far more accurate metric \n",
    "from sklearn.model_selection import GridSearchCV\n",
    "param_grid = {\n",
    "    'bootstrap': [False],\n",
    "    'max_depth': [65, 75, 85, 95], \n",
    "    'max_features': [2,3],\n",
    "    'min_samples_leaf': [1,2,3],\n",
    "    'min_samples_split': [4,5,6],\n",
    "    'n_estimators': [400,500, 550, 650]\n",
    "}\n",
    "\n",
    "rf = RandomForestClassifier() \n",
    "grid_search = GridSearchCV(estimator = rf, param_grid = param_grid , cv = 3, n_jobs = -1, verbose = 2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 288 candidates, totalling 864 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  25 tasks      | elapsed:   15.0s\n",
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/joblib/externals/loky/process_executor.py:706: UserWarning: A worker stopped while some jobs were given to the executor. This can be caused by a too short worker timeout or by a memory leak.\n",
      "  \"timeout or by a memory leak.\", UserWarning\n",
      "[Parallel(n_jobs=-1)]: Done 146 tasks      | elapsed:   50.9s\n",
      "[Parallel(n_jobs=-1)]: Done 349 tasks      | elapsed:  1.9min\n",
      "[Parallel(n_jobs=-1)]: Done 632 tasks      | elapsed:  3.4min\n",
      "[Parallel(n_jobs=-1)]: Done 864 out of 864 | elapsed:  4.6min finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=3, error_score='raise-deprecating',\n",
       "             estimator=RandomForestClassifier(bootstrap=True, class_weight=None,\n",
       "                                              criterion='gini', max_depth=None,\n",
       "                                              max_features='auto',\n",
       "                                              max_leaf_nodes=None,\n",
       "                                              min_impurity_decrease=0.0,\n",
       "                                              min_impurity_split=None,\n",
       "                                              min_samples_leaf=1,\n",
       "                                              min_samples_split=2,\n",
       "                                              min_weight_fraction_leaf=0.0,\n",
       "                                              n_estimators='warn', n_jobs=None,\n",
       "                                              oob_score=False,\n",
       "                                              random_state=None, verbose=0,\n",
       "                                              warm_start=False),\n",
       "             iid='warn', n_jobs=-1,\n",
       "             param_grid={'bootstrap': [False], 'max_depth': [65, 75, 85, 95],\n",
       "                         'max_features': [2, 3], 'min_samples_leaf': [1, 2, 3],\n",
       "                         'min_samples_split': [4, 5, 6],\n",
       "                         'n_estimators': [400, 500, 550, 650]},\n",
       "             pre_dispatch='2*n_jobs', refit=True, return_train_score=False,\n",
       "             scoring=None, verbose=2)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "grid_search.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'bootstrap': False, 'max_depth': 65, 'max_features': 3, 'min_samples_leaf': 2, 'min_samples_split': 4, 'n_estimators': 400}\n"
     ]
    }
   ],
   "source": [
    "print(grid_search.best_params_) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate(model, test_features, test_labels):\n",
    "    predictions = model.predict(test_features)\n",
    "    vals = abs(predictions - test_labels)\n",
    "    errors = [] \n",
    "    for x in vals:\n",
    "        if x == 1:\n",
    "            errors.append(x)\n",
    "    print(errors)\n",
    "    accuracy = 100 - (100 * ( len(errors) / len(predictions)))\n",
    "    print('Model Performance')\n",
    "    print('Accuracy = {:0.2f}%.'.format(accuracy))\n",
    "    \n",
    "    return accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/danielraad/anaconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
      "  \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]\n",
      "Model Performance\n",
      "Accuracy = 94.16%.\n",
      "[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]\n",
      "Model Performance\n",
      "Accuracy = 96.35%.\n",
      "Improvement of 2.33%.\n"
     ]
    }
   ],
   "source": [
    "base_model = RandomForestClassifier()\n",
    "base_model.fit(X_train, y_train)\n",
    "base_accuracy = evaluate(base_model, X_test, y_tests)\n",
    "np.seterr('raise')\n",
    "\n",
    "best_model1 = RandomForestClassifier(bootstrap = False, max_depth = 65, max_features = 3, min_samples_leaf = 2, min_samples_split = 4, n_estimators = 400)\n",
    "best_model1.fit(X_train, y_train)\n",
    "best_accuracy1 = evaluate (best_model1, X_test, y_tests)\n",
    "# best_model = grid_search.best_estimator_\n",
    "# best_accuracy = evaluate(best_model, X_test, y_tests)\n",
    "\n",
    "print('Improvement of {:0.2f}%.'.format( 100 * (best_accuracy1 - base_accuracy) / base_accuracy))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9451186564439575\n"
     ]
    }
   ],
   "source": [
    "#Scoring with kfold \n",
    "from sklearn.model_selection import cross_val_score\n",
    "scores = cross_val_score(best_model1,df,y_red,cv=10,scoring='accuracy')\n",
    "print(scores.mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Evaluation\n",
    "from sklearn.metrics import confusion_matrix\n",
    "import itertools\n",
    "\n",
    "def plot_confusion_matrix(cm, classes,\n",
    "                          normalize=False,\n",
    "                          title='Confusion matrix',\n",
    "                          cmap=plt.cm.Oranges):\n",
    "    \"\"\"\n",
    "    This function prints and plots the confusion matrix.\n",
    "    Normalization can be applied by setting `normalize=True`.\n",
    "    Source: http://scikit-learn.org/stable/auto_examples/model_selection/plot_confusion_matrix.html\n",
    "    \"\"\"\n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "        print(\"Normalized confusion matrix\")\n",
    "    else:\n",
    "        print('Confusion matrix, without normalization')\n",
    "\n",
    "    print(cm)\n",
    "\n",
    "    plt.figure(figsize = (10, 10))\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title, size = 24)\n",
    "    plt.colorbar(aspect=4)\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation=45, size = 14)\n",
    "    plt.yticks(tick_marks, classes, size = 14)\n",
    "\n",
    "    fmt = '.2f' if normalize else 'd'\n",
    "    thresh = cm.max() / 2.\n",
    "    \n",
    "    # Labeling the plot\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, format(cm[i, j], fmt), fontsize = 20,\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "        \n",
    "    plt.grid(None)\n",
    "    plt.tight_layout()\n",
    "    plt.ylabel('True label', size = 18)\n",
    "    plt.xlabel('Predicted label', size = 18)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n",
      "[[180  10]\n",
      " [ 10 348]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Cancer confusion matrix for the pruned model, finding the number of errors made on the test set \n",
    "rf_predictions = best_model1.predict(X_test)\n",
    "cm = confusion_matrix(y_tests, rf_predictions)\n",
    "plot_confusion_matrix(cm, classes = ['No BRCA', 'BRCA'],\n",
    "                      title = 'Cancer Confusion Matrix')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate_model(predictions, probs, train_predictions, train_probs):\n",
    "    \"\"\"Compare machine learning model to baseline performance.\n",
    "    Computes statistics and shows ROC curve.\"\"\"\n",
    "    \n",
    "    baseline = {}\n",
    "    \n",
    "    baseline['recall'] = recall_score(y_tests, [1 for _ in range(len(y_tests))])\n",
    "    baseline['precision'] = precision_score(y_tests, [1 for _ in range(len(y_tests))])\n",
    "    baseline['roc'] = 0.5\n",
    "    \n",
    "    results = {}\n",
    "    \n",
    "    results['recall'] = recall_score(y_tests, predictions)\n",
    "    results['precision'] = precision_score(y_tests, predictions)\n",
    "    results['roc'] = roc_auc_score(y_tests, probs)\n",
    "    \n",
    "    train_results = {}\n",
    "    train_results['recall'] = recall_score(y_train, train_predictions)\n",
    "    train_results['precision'] = precision_score(y_train, train_predictions)\n",
    "    train_results['roc'] = roc_auc_score(y_train, train_probs)\n",
    "    \n",
    "    for metric in ['recall', 'precision', 'roc']:\n",
    "        print(f'{metric.capitalize()} Baseline: {round(baseline[metric], 2)} Test: {round(results[metric], 2)} Train: {round(train_results[metric], 2)}')\n",
    "    \n",
    "    # Calculate false positive rates and true positive rates\n",
    "    base_fpr, base_tpr, _ = roc_curve(y_tests, [1 for _ in range(len(y_tests))])\n",
    "    model_fpr, model_tpr, _ = roc_curve(y_tests, probs)\n",
    "\n",
    "    plt.figure(figsize = (8, 6))\n",
    "    plt.rcParams['font.size'] = 16\n",
    "    \n",
    "    # Plot both curves\n",
    "    plt.plot(base_fpr, base_tpr, 'b', label = 'baseline')\n",
    "    plt.plot(model_fpr, model_tpr, 'r', label = 'model')\n",
    "    plt.legend();\n",
    "    plt.xlabel('False Positive Rate'); plt.ylabel('True Positive Rate'); plt.title('ROC Curves');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Recall Baseline: 1.0 Test: 0.97 Train: 1.0\n",
      "Precision Baseline: 0.65 Test: 0.97 Train: 1.0\n",
      "Roc Baseline: 0.5 Test: 0.99 Train: 1.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "##AUC ROC testing for the pruned model \n",
    "\n",
    "train_rf_predictions = best_model1.predict(X_train)\n",
    "train_rf_probs = best_model1.predict_proba(X_train)[:,1]\n",
    "rf_predictions = best_model1.predict(X_test)\n",
    "rf_probs = best_model1.predict_proba(X_test)[:,1]\n",
    "evaluate_model(rf_predictions, rf_probs, train_rf_predictions, train_rf_probs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "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>feature</th>\n",
       "      <th>importance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12542</th>\n",
       "      <td>NARFL</td>\n",
       "      <td>0.001615</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12425</th>\n",
       "      <td>MYO18A</td>\n",
       "      <td>0.001239</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23974</th>\n",
       "      <td>ZNF555</td>\n",
       "      <td>0.001139</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6400</th>\n",
       "      <td>FRG2</td>\n",
       "      <td>0.001133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23584</th>\n",
       "      <td>ZDHHC19</td>\n",
       "      <td>0.001128</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23975</th>\n",
       "      <td>ZNF556</td>\n",
       "      <td>0.001104</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10331</th>\n",
       "      <td>MATK</td>\n",
       "      <td>0.001101</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13164</th>\n",
       "      <td>NUDT16L1</td>\n",
       "      <td>0.001093</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17216</th>\n",
       "      <td>RN7SL91P</td>\n",
       "      <td>0.001093</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22535</th>\n",
       "      <td>TXNDC11</td>\n",
       "      <td>0.001057</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        feature  importance\n",
       "12542     NARFL    0.001615\n",
       "12425    MYO18A    0.001239\n",
       "23974    ZNF555    0.001139\n",
       "6400       FRG2    0.001133\n",
       "23584   ZDHHC19    0.001128\n",
       "23975    ZNF556    0.001104\n",
       "10331      MATK    0.001101\n",
       "13164  NUDT16L1    0.001093\n",
       "17216  RN7SL91P    0.001093\n",
       "22535   TXNDC11    0.001057"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#The most important features \n",
    "fi_model = pd.DataFrame({'feature': features,\n",
    "                   'importance': best_model1.feature_importances_}).\\\n",
    "                    sort_values('importance', ascending = False)\n",
    "fi_model.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "##Doing a second grid search for more accurate values \n",
    "##"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 108 candidates, totalling 324 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  39 tasks      | elapsed:   32.5s\n",
      "[Parallel(n_jobs=-1)]: Done 160 tasks      | elapsed:  1.9min\n",
      "[Parallel(n_jobs=-1)]: Done 324 out of 324 | elapsed:  3.6min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'bootstrap': False, 'max_depth': 65, 'max_features': 3, 'min_samples_leaf': 2, 'min_samples_split': 4, 'n_estimators': 450}\n"
     ]
    }
   ],
   "source": [
    "#this provides far more accurate metric \n",
    "from sklearn.model_selection import GridSearchCV\n",
    "param_grid2 = {\n",
    "    'bootstrap': [False],\n",
    "    'max_depth': [55, 60, 65, 70], \n",
    "    'max_features': [2,3,4],\n",
    "    'min_samples_leaf': [2],\n",
    "    'min_samples_split': [3,4,5],\n",
    "    'n_estimators': [350, 400, 450]\n",
    "}\n",
    "\n",
    "rf = RandomForestClassifier() \n",
    "grid_search2 = GridSearchCV(estimator = rf, param_grid = param_grid2 , cv = 3, n_jobs = -1, verbose = 2, pre_dispatch = 2 )\n",
    "grid_search2.fit(X_train, y_train)\n",
    "print(grid_search2.best_params_) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 108 candidates, totalling 324 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  39 tasks      | elapsed:   23.0s\n",
      "[Parallel(n_jobs=-1)]: Done 160 tasks      | elapsed:  1.7min\n",
      "[Parallel(n_jobs=-1)]: Done 324 out of 324 | elapsed:  3.5min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'bootstrap': False, 'max_depth': 55, 'max_features': 4, 'min_samples_leaf': 2, 'min_samples_split': 5, 'n_estimators': 450}\n"
     ]
    }
   ],
   "source": [
    "#this provides far more accurate metric \n",
    "from sklearn.model_selection import GridSearchCV\n",
    "param_grid3 = {\n",
    "    'bootstrap': [False],\n",
    "    'max_depth': [65], \n",
    "    'max_features': [3],\n",
    "    'min_samples_leaf': [2],\n",
    "    'min_samples_split': [4],\n",
    "    'n_estimators': [425, 450, 475]\n",
    "}\n",
    "\n",
    "rf = RandomForestClassifier() \n",
    "grid_search3 = GridSearchCV(estimator = rf, param_grid = param_grid2 , cv = 3, n_jobs = -1, verbose = 2, pre_dispatch = 2)\n",
    "grid_search3.fit(X_train, y_train)\n",
    "print(grid_search3.best_params_) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]\n",
      "Model Performance\n",
      "Accuracy = 96.35%.\n",
      "Improvement of 0.00%.\n"
     ]
    }
   ],
   "source": [
    "final_model = grid_search3.best_estimator_\n",
    "final_accuracy = evaluate (final_model, X_test, y_tests)\n",
    "\n",
    "print('Improvement of {:0.2f}%.'.format( 100 * (final_accuracy - best_accuracy1) / best_accuracy1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9439101861993429\n"
     ]
    }
   ],
   "source": [
    "#Scoring with kfold \n",
    "from sklearn.model_selection import cross_val_score\n",
    "scores = cross_val_score(final_model,df,y_red,cv=10,scoring='accuracy')\n",
    "print(scores.mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}