random_forest.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Random Forest makes prediction based on individual outputs of many decision trees. The goal is to improve upon the original foreast that used a single decision tree and achieved an accuracy around 0.934."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Read the data\n",
    "X_train = pd.read_csv('X_train_w_ID.csv', index_col=0)\n",
    "X_test = pd.read_csv('X_test_w_ID.csv', index_col = 0)\n",
    "Y_train = pd.read_csv('Y_train.csv', index_col = 0, header= None, names = ['Y'])\n",
    "Y_test = pd.read_csv('Y_test.csv', index_col = 0, header= None, names = ['Y'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Remove patient ID from the training data\n",
    "x_train = X_train.drop(\"DESYNPUF_ID\",axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "83270"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train.shape[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "27"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_test.shape[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Remove patient ID from testing set\n",
    "x_test = X_test.drop('DESYNPUF_ID',axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/ksenia/anaconda/envs/py36/lib/python3.6/site-packages/ipykernel_launcher.py:6: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples,), for example using ravel().\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "\n",
    "#Build a random forest using 100 estimators for testing purposes\n",
    "rf_def = RandomForestClassifier(criterion='entropy', n_estimators=100)\n",
    "rf_def = rf_def.fit(x_train, Y_train)\n",
    "rf_pred = rf_def.predict_proba(x_test)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Function to print ROC curves. \n",
    "from sklearn.metrics import roc_curve, auc\n",
    "\n",
    "def plotUnivariateROC(preds, truth, label_string):\n",
    "    '''\n",
    "    preds is an nx1 array of predictions\n",
    "    truth is an nx1 array of truth labels\n",
    "    label_string is text to go into the plotting label\n",
    "    '''\n",
    "    #Student input code here\n",
    "    #1. call the roc_curve function to get the ROC X and Y values\n",
    "    fpr, tpr, thresholds = roc_curve(truth, preds)\n",
    "    #2. Input fpr and tpr into the auc function to get the AUC\n",
    "    roc_auc = auc(fpr, tpr)\n",
    "    \n",
    "    #we are doing this as a special case because we are sending unfitted predictions\n",
    "    #into the function\n",
    "    if roc_auc < 0.5:\n",
    "        fpr, tpr, thresholds = roc_curve(truth, -1 * preds)\n",
    "        roc_auc = auc(fpr, tpr)\n",
    "\n",
    "    #chooses a random color for plotting\n",
    "    c = (np.random.rand(), np.random.rand(), np.random.rand())\n",
    "\n",
    "    #create a plot and set some options\n",
    "    plt.plot(fpr, tpr, color = c, label = label_string + ' (AUC = %0.3f)' % roc_auc)\n",
    "    \n",
    "\n",
    "    plt.plot([0, 1], [0, 1], 'k--')\n",
    "    plt.xlim([0.0, 1.0])\n",
    "    plt.ylim([0.0, 1.0])\n",
    "    plt.xlabel('FPR')\n",
    "    plt.ylabel('TPR')\n",
    "    plt.title('ROC')\n",
    "    plt.legend(loc=\"lower right\")\n",
    "    \n",
    "    return roc_auc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYoAAAEWCAYAAAB42tAoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nzQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1109d12e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_curve, auc, roc_auc_score\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plotUnivariateROC(rf_pred[:,1], Y_test, 'RF')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "A simple Random Forest already outperformed the original decision tree forecast by producing an AUC of 0.953."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next I will use use out-of-bag error to tune parameters for the random forest and further improve accuracy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYoAAAEWCAYAAAB42tAoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x110f7de10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#This is just to show that OOB method produces similar results \n",
    "rf_def = RandomForestClassifier(criterion='entropy', n_estimators=100, oob_score=True)\n",
    "rf_def = rf_def.fit(x_train, Y_train.values.ravel())\n",
    "rf_pred_test = rf_def.predict_proba(x_test)\n",
    "rf_pred_train = rf_def.predict_proba(x_train)\n",
    "\n",
    "plotUnivariateROC(rf_pred_test[:,1], Y_test, 'Test')\n",
    "plotUnivariateROC( rf_pred_train[:,1],Y_train, 'Train')\n",
    "plotUnivariateROC(rf_def.oob_decision_function_[:,1], Y_train, 'OOB')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now use OOB to perform model selection."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Select different estimator sizes and number of features\n",
    "n_est = [50, 100, 200, 500, 1000]\n",
    "m_feat = [1, 3, 6, 11]\n",
    "\n",
    "aucs_oob = {}\n",
    "aucs_test = {}\n",
    "\n",
    "for m in m_feat:\n",
    "    aucs_oob[m] = []\n",
    "    aucs_test[m] = []\n",
    "    for n in n_est:\n",
    "        rf_oob = RandomForestClassifier(criterion='entropy', n_estimators=n, max_features=m, oob_score=True)\n",
    "        rf_oob = rf_oob.fit(x_train, Y_train.values.ravel())\n",
    "        aucs_oob[m].append(roc_auc_score(Y_train, rf_oob.oob_decision_function_[:,1]))\n",
    "        aucs_test[m].append(roc_auc_score(Y_test, rf_oob.predict_proba(x_test)[:,1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYwAAAEWCAYAAAB1xKBvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCCGEqkSHEUIIoSr/AzGY6KGQ2SeMAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1110cb4e0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plot results\n",
    "\n",
    "x = np.log2(np.array(n_est))\n",
    "for m in m_feat:\n",
    "    plt.plot(x, aucs_oob[m], label='max_feat={}'.format(m))\n",
    "    \n",
    "plt.title('OOB AUC by Max Feat and N-Estimators')\n",
    "plt.xlabel('Log2(N-Estimators)')\n",
    "plt.ylabel('OOB-AUC')\n",
    "plt.legend(loc=4, ncol=2, prop={'size':10})\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And model with 11 max features and 1000 estimators produced the highest accuracy.\n",
    "\n",
    "Based on the chart, further increasing number of estimators might slowly continue improving the results, but the improvement may be only marginal. On the other hand, increasing number of features may be more promising."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Build a random forest with 12 max features and 1000 estimators.\n",
    "big_rf = RandomForestClassifier(criterion='entropy', max_features = 12, n_estimators=1000)\n",
    "big_rf = big_rf.fit(x_train, Y_train.values.ravel())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 0.999951963492 \n",
      "Test accuracy: 0.948409410239 \n",
      "Difference: 0.0515425532534\n",
      "\n",
      "Train recall:  0.999536446865 \n",
      "Test recall:  0.572705397044 \n",
      "Difference:  0.426831049822\n"
     ]
    }
   ],
   "source": [
    "#Calculate accuracy and recall\n",
    "from sklearn import metrics\n",
    "\n",
    "#Predicting train_df\n",
    "fcst_train = big_rf.predict(x_train)\n",
    "\n",
    "#Predicting test_df\n",
    "fcst_test = big_rf.predict(x_test)\n",
    "\n",
    "train_accuracy = big_rf.score(x_train, Y_train)\n",
    "test_accuracy = big_rf.score(x_test, Y_test)\n",
    "\n",
    "train_recall = metrics.recall_score(Y_train, fcst_train)\n",
    "test_recall = metrics.recall_score(Y_test, fcst_test)\n",
    "\n",
    "print('Train accuracy:',train_accuracy, '\\nTest accuracy:', test_accuracy, '\\nDifference:', \n",
    "     train_accuracy - test_accuracy)\n",
    "\n",
    "print('\\nTrain recall: ', train_recall, '\\nTest recall: ', test_recall, '\\nDifference: ', train_recall-test_recall)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 1, 1, ..., 1, 1, 1])"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample_weight = np.array([1 if i == 0 else 9 for i in Y_train.values])\n",
    "\n",
    "sample_weight"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The Recall performance was below the simple Decision Tree we tested previously which produced a 70% recall. Now I'll retrain the Random Forest to determine the best parameters for maximizing Recall instead of accuracy.\n",
    "\n",
    "I'll also try rebalancing classes to reflect on higher prevalence of negative target in the dataset for improving Recall."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Select different estimator sizes and number of features\n",
    "n_est = [50, 100, 200, 500, 1000]\n",
    "m_feat = [1, 3, 6, 11]\n",
    "\n",
    "recall_oob = {}\n",
    "recall_test = {}\n",
    "\n",
    "sample_weight = np.array([1 if i == 0 else 9 for i in Y_train.values])\n",
    "\n",
    "for m in m_feat:\n",
    "    recall_oob[m] = []\n",
    "    recall_test[m] = []\n",
    "    for n in n_est:\n",
    "        rf_oob = RandomForestClassifier(criterion='entropy', n_estimators=n, max_features=m, oob_score=True)\n",
    "        rf_oob = rf_oob.fit(x_train, Y_train.values.ravel(), sample_weight=sample_weight)\n",
    "        recall_oob[m].append(metrics.recall_score(Y_train, rf_oob.predict(x_train)))\n",
    "        recall_test[m].append(metrics.recall_score(Y_test, rf_oob.predict(x_test)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZgAAAEWCAYAAABbgYH9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x113f99940>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plot results\n",
    "\n",
    "x = np.log2(np.array(n_est))\n",
    "for m in m_feat:\n",
    "    plt.plot(x, recall_oob[m], label='max_feat={}'.format(m))\n",
    "    \n",
    "plt.title('OOB Recall by Max Feat and N-Estimators')\n",
    "plt.xlabel('Log2(N-Estimators)')\n",
    "plt.ylabel('OOB-Recall')\n",
    "plt.legend(loc=4, ncol=2, prop={'size':10})\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nic8pAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x112a4dfd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "x = np.log2(np.array(n_est))\n",
    "for m in m_feat:\n",
    "    plt.plot(x, recall_test[m], label='max_feat={}'.format(m))\n",
    "    \n",
    "plt.title('Test Recall by Max Feat and N-Estimators')\n",
    "plt.xlabel('Log2(N-Estimators)')\n",
    "plt.ylabel('Recall')\n",
    "plt.legend(loc=4, ncol=2, prop={'size':10})\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Build a random forest based on the parameter tuning above\n",
    "big_rf = RandomForestClassifier(criterion='entropy', max_features = None, n_estimators=100)\n",
    "big_rf = big_rf.fit(x_train, Y_train.values.ravel())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 0.999927945238 \n",
      "Test accuracy: 0.959181467738 \n",
      "Difference: 0.0407464775005\n",
      "\n",
      "Train recall:  0.999304670298 \n",
      "Test recall:  0.675833619801 \n",
      "Difference:  0.323471050497\n"
     ]
    }
   ],
   "source": [
    "#Calculate accuracy and recall\n",
    "from sklearn import metrics\n",
    "\n",
    "#Predicting train_df\n",
    "fcst_train = big_rf.predict(x_train)\n",
    "\n",
    "#Predicting test_df\n",
    "fcst_test = big_rf.predict(x_test)\n",
    "\n",
    "train_accuracy = big_rf.score(x_train, Y_train)\n",
    "test_accuracy = big_rf.score(x_test, Y_test)\n",
    "\n",
    "train_recall = metrics.recall_score(Y_train, fcst_train)\n",
    "test_recall = metrics.recall_score(Y_test, fcst_test)\n",
    "\n",
    "print('Train accuracy:',train_accuracy, '\\nTest accuracy:', test_accuracy, '\\nDifference:', \n",
    "     train_accuracy - test_accuracy)\n",
    "\n",
    "print('\\nTrain recall: ', train_recall, '\\nTest recall: ', test_recall, '\\nDifference: ', train_recall-test_recall)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Setting max_teatures to None improved recall. I will perform grid search next to further tune the parameters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.grid_search import GridSearchCV\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "rfc = RandomForestClassifier(criterion='entropy', max_features= 'sqrt', n_estimators=200, oob_score = True) \n",
    "\n",
    "param_grid = { \n",
    "    'max_depth' : [30, 100, None],\n",
    "    'n_estimators': [10, 20, 50, 200],\n",
    "    'max_features': [None, 10, 20, 'auto'],\n",
    "    'min_samples_leaf' : [2, 50, 100]   \n",
    "}\n",
    "\n",
    "CV_rfc = GridSearchCV(estimator=rfc, scoring='recall', param_grid=param_grid, cv= 5)\n",
    "CV_rfc.fit(x_train, Y_train.values.ravel())\n",
    "\n",
    "print(CV_rfc.best_params_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Build a random forest based on the parameter tuning above\n",
    "rf = RandomForestClassifier(criterion='entropy', max_features = None, n_estimators=50, min_samples_leaf=2)\n",
    "rf = rf.fit(x_train, Y_train.values.ravel())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Final Version of Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy: 0.998186621833 \n",
      "Test accuracy: 0.958424901827 \n",
      "Difference: 0.039761720006\n",
      "\n",
      "Train recall:  0.983427975432 \n",
      "Test recall:  0.669645926435 \n",
      "Difference:  0.313782048996\n",
      "\n",
      "Train precision: 0.9990581587 \n",
      "Test precision: 0.909855207847 \n",
      "Difference: 0.0892029508535\n",
      "\n",
      "Test AUC:  0.980369916884\n"
     ]
    }
   ],
   "source": [
    "#Calculate accuracy and recall\n",
    "from sklearn import metrics\n",
    "\n",
    "#Predicting train_df\n",
    "fcst_train = rf.predict(x_train)\n",
    "\n",
    "#Predicting test_df\n",
    "fcst_test = rf.predict(x_test)\n",
    "\n",
    "train_accuracy = rf.score(x_train, Y_train)\n",
    "test_accuracy = rf.score(x_test, Y_test)\n",
    "\n",
    "train_recall = metrics.recall_score(Y_train, fcst_train)\n",
    "test_recall = metrics.recall_score(Y_test, fcst_test)\n",
    "\n",
    "train_precision = metrics.precision_score(Y_train, fcst_train)\n",
    "test_precision = metrics.precision_score(Y_test, fcst_test)\n",
    "\n",
    "test_auc = roc_auc_score(Y_test, rf.predict_proba(x_test)[:,1])\n",
    "\n",
    "print('Train accuracy:',train_accuracy, '\\nTest accuracy:', test_accuracy, '\\nDifference:', \n",
    "     train_accuracy - test_accuracy)\n",
    "\n",
    "print('\\nTrain recall: ', train_recall, '\\nTest recall: ', test_recall, '\\nDifference: ', train_recall-test_recall)\n",
    "\n",
    "print('\\nTrain precision:',train_precision, '\\nTest precision:', test_precision, '\\nDifference:', \n",
    "     train_precision - test_precision)\n",
    "\n",
    "print('\\nTest AUC: ', test_auc)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.914804469274\n"
     ]
    }
   ],
   "source": [
    "test_precision = metrics.precision_score(Y_test, fcst_test)\n",
    "print(test_precision)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0.00389311  0.00355746  0.02513415  0.0056863   0.00751837  0.09992779\n",
      "  0.00489253  0.0283415   0.00567214  0.02004541  0.02479427  0.00370463\n",
      "  0.00305991  0.00557048  0.17825189  0.00245743  0.04797436  0.09715148\n",
      "  0.02454827  0.02955498  0.03638145  0.1211047   0.02940973  0.12360755\n",
      "  0.03199679  0.02702181  0.00874151]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABJEAAALUCAYAAAC7C6NXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x116f97ba8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Visualize feature importance\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "importances = rf.feature_importances_\n",
    "#print(len(importances))\n",
    "\n",
    "labels = x_train.columns\n",
    "\n",
    "assert(len(importances) == len(labels))\n",
    "\n",
    "print(importances)\n",
    "\n",
    "#Sort and order importances in descending order\n",
    "importances.sort()\n",
    "importances = importances[::-1]\n",
    "\n",
    "N = len(importances)\n",
    "xx = np.linspace(0, 1, N)\n",
    "plt.figure(figsize=(20,10))\n",
    "plt.bar(xx, importances, width=1/N/1.5)\n",
    "plt.xticks(xx, labels, rotation='vertical')\n",
    "plt.title('Features by importance')\n",
    "\n",
    "plt.grid(True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Function to visualize classification\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "def visualize_classifier(model, X, y, ax=None, cmap='rainbow'):\n",
    "    ax = ax or plt.gca()\n",
    "    \n",
    "    # Plot the training points\n",
    "    ax.scatter(X[:, 0], X[:, 1], c=y, s=30, cmap=cmap,\n",
    "               clim=(y.min(), y.max()), zorder=3)\n",
    "    ax.axis('tight')\n",
    "    ax.axis('off')\n",
    "    xlim = ax.get_xlim()\n",
    "    ylim = ax.get_ylim()\n",
    "    \n",
    "    # fit the estimator\n",
    "    model.fit(X, y)\n",
    "    xx, yy = np.meshgrid(np.linspace(*xlim, num=200),\n",
    "                         np.linspace(*ylim, num=200))\n",
    "    Z = model.predict(np.c_[xx.ravel(), yy.ravel()]).reshape(xx.shape)\n",
    "\n",
    "    # Create a color plot with the results\n",
    "    n_classes = len(np.unique(y))\n",
    "    contours = ax.contourf(xx, yy, Z, alpha=0.3,\n",
    "                           levels=np.arange(n_classes + 1) - 0.5,\n",
    "                           cmap=cmap, clim=(y.min(), y.max()),\n",
    "                           zorder=1)\n",
    "\n",
    "    ax.set(xlim=xlim, ylim=ylim)\n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "rf = RandomForestClassifier(criterion='entropy', max_features = None, n_estimators=50, min_samples_leaf=2)\n",
    "X = x_train[['unique_ip_dt_mm', '#Rx_mm']].as_matrix()\n",
    "y = Y_train.as_matrix()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_test.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "visualize_classifier(rf, X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}