classifier.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/devaraj/anaconda/lib/python2.7/site-packages/IPython/kernel/__main__.py:22: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy\n"
     ]
    }
   ],
   "source": [
    "# Read and clean the data\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd \n",
    "import numpy as np \n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "\n",
    "# Label encoding for string values\n",
    "def LabelEncoding(col):\n",
    "        lbl_enc = LabelEncoder()\n",
    "        lbl_enc.fit(col)\n",
    "        cat_col = lbl_enc.transform(col)\n",
    "        return cat_col\n",
    "\n",
    "def clean_my_df(df):\n",
    "\n",
    "    numeric_features = df._get_numeric_data()\n",
    "    \n",
    "    # Normalize the scale of numeric values. \n",
    "    numeric_features = (numeric_features) / (numeric_features.max())\n",
    "    categorical_features = df.select_dtypes(include=['object'])\n",
    "\n",
    "    for col_name in categorical_features.columns :\n",
    "        categorical_features[col_name] = categorical_features[col_name].astype(\"category\",\n",
    "                                        categories=pd.unique(categorical_features[col_name].values.ravel()))\n",
    "        data_oneHot = pd.get_dummies(categorical_features[col_name]) # One hot encoding\n",
    "        numeric_features = pd.concat([numeric_features,data_oneHot],axis = 1)\n",
    "\n",
    "    return numeric_features\n",
    "\n",
    "\n",
    "data = pd.read_csv(\"data.csv\")\n",
    "data.drop([\"Customer\"],axis = 1, inplace = True)\n",
    "data.drop([\"Effective To Date\"],axis = 1, inplace = True)\n",
    "labels = pd.DataFrame({\"Response\":LabelEncoding(data[\"Response\"])})\n",
    "data.drop([\"Response\"],axis = 1, inplace = True)\n",
    "features = clean_my_df(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABQ0AAAR4CAYAAACsHPBfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nnf7MAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2401a12750>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Feature Importances\n",
    "import xgboost as xgb\n",
    "from collections import OrderedDict\n",
    "\n",
    "# Check feature importances for the classifier.\n",
    "xgb_params = {\"objective\": \"binary:logistic\",\"max_depth\": 8}\n",
    "num_rounds = 1000\n",
    "xg_train = xgb.DMatrix(features,label=labels)\n",
    "bst = xgb.train(xgb_params, xg_train, num_rounds)\n",
    "importances = bst.get_fscore()\n",
    "importances = OrderedDict(sorted(importances.items(), key=lambda x: x[1]))\n",
    "\n",
    "# Plot the importan features with fscore\n",
    "plt.figure(figsize=(20,20))\n",
    "plt.barh(range(len(importances)), importances.values(), align='center')\n",
    "plt.yticks(range(len(importances)),importances.keys())\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index([u'Urban', u'Large', u'Divorced', u'Small', u'Suburban', u'Call Center',\n",
      "       u'SUV', u'High School or Below', u'Branch', u'Extended', u'Bachelor',\n",
      "       u'Single', u'Retired', u'Basic', u'Four-Door Car', u'Offer1',\n",
      "       u'Medsize', u'College', u'Married', u'Number of Open Complaints',\n",
      "       u'Agent', u'Offer2', u'F', u'Number of Policies',\n",
      "       u'Monthly Premium Auto', u'Months Since Last Claim',\n",
      "       u'Total Claim Amount', u'Months Since Policy Inception',\n",
      "       u'Customer Lifetime Value', u'Income'],\n",
      "      dtype='object')\n"
     ]
    }
   ],
   "source": [
    "# Important Features\n",
    "imp_features = features[importances.keys()[-30:]]\n",
    "print imp_features.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "score: 0.897865669842 \n",
      "\n",
      "{'colsample_bytree': 0.9,\n",
      " 'learning_rate': 0.1,\n",
      " 'max_depth': 9,\n",
      " 'min_child_weight': 5,\n",
      " 'n_estimators': 10,\n",
      " 'nthread': 4,\n",
      " 'objective': 'binary:logistic',\n",
      " 'subsample': 0.75}\n"
     ]
    }
   ],
   "source": [
    "# Parameter Tuning\n",
    "from sklearn.cross_validation import *\n",
    "from sklearn.metrics import *\n",
    "from sklearn.grid_search import GridSearchCV\n",
    "from pprint import pprint\n",
    "\n",
    "# List of hyperparametres\n",
    "xgb_model = xgb.XGBClassifier()\n",
    "parameters = {'nthread':[4],\n",
    "              'objective':['binary:logistic'],\n",
    "              'learning_rate': [0.01,0.1],\n",
    "              'max_depth': [6,9],\n",
    "              'min_child_weight': [5,8],\n",
    "              'subsample': [0.6,0.75],\n",
    "              'colsample_bytree': [0.75,0.9],\n",
    "              'n_estimators': [10]\n",
    "              }\n",
    "\n",
    "# Scoring function for evaluation\n",
    "def score_func(model, X, y_true) :\n",
    "    preds = model.predict(X)\n",
    "    return accuracy_score(y_true,preds)\n",
    "\n",
    "# This step is used as Gridsearch does not handle well with nupy arrays\n",
    "labels = np.array([l[0] for l in labels.values.tolist()])\n",
    "\n",
    "# Grid Search using stratified K fold\n",
    "clf = GridSearchCV(xgb_model,parameters,n_jobs=4,cv=StratifiedKFold(labels,n_folds=4, shuffle=True),\n",
    "                   scoring=score_func,refit=True)\n",
    "clf.fit(imp_features,labels)\n",
    "best_parameters,score,filler = max(clf.grid_scores_,key=lambda x:x[1])\n",
    "\n",
    "#Get Best hyper parameters\n",
    "print \"score:\", score, \"\\n\"\n",
    "pprint(best_parameters)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "Note : Parameter tuning can be improved by taking more values of the hyperparameters, but it takes too much time. Therefore using only few values to demonstarte the procedure. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "          0       0.91      0.99      0.95      1788\n",
      "          1       0.90      0.39      0.55       286\n",
      "\n",
      "avg / total       0.91      0.91      0.89      2074\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#classification Report\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.cross_validation import train_test_split\n",
    "\n",
    "# Split the files into training and testing\n",
    "x_train, x_test, y_train, y_test = train_test_split(imp_features,labels)\n",
    "\n",
    "d_train = xgb.DMatrix(x_train, label=y_train)\n",
    "d_test = xgb.DMatrix(x_test)\n",
    "\n",
    "# Use best parameters from hyperparameter tuning\n",
    "param = best_parameters\n",
    "param['n_estimators'] = 10\n",
    "\n",
    "bst = xgb.train(param,d_train)\n",
    "pred = bst.predict(d_test)\n",
    "pred_transformed = []\n",
    "\n",
    "# Transform the probabilites to class labels.\n",
    "for i in range(len(pred)) : \n",
    "    if pred[i] > 0.5 :\n",
    "        pred_transformed.append(1)\n",
    "    else :\n",
    "        pred_transformed.append(0)\n",
    "\n",
    "#Get Classification Report\n",
    "pred_transformed = np.array(pred_transformed)\n",
    "print(classification_report(y_test,pred_transformed))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The recall of 1 (i.e response YES) is low due to less number of samples of that class. "
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}