Taiwan credit default machine learning.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.grid_search import GridSearchCV\n",
    "from sklearn.cross_validation import train_test_split\n",
    "from sklearn.metrics import confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(30000, 25)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = pd.read_excel(\"default of credit card clients.xls\", skiprows = 1)\n",
    "data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ID</th>\n",
       "      <th>LIMIT_BAL</th>\n",
       "      <th>SEX</th>\n",
       "      <th>EDUCATION</th>\n",
       "      <th>MARRIAGE</th>\n",
       "      <th>AGE</th>\n",
       "      <th>PAY_0</th>\n",
       "      <th>PAY_2</th>\n",
       "      <th>PAY_3</th>\n",
       "      <th>PAY_4</th>\n",
       "      <th>...</th>\n",
       "      <th>BILL_AMT4</th>\n",
       "      <th>BILL_AMT5</th>\n",
       "      <th>BILL_AMT6</th>\n",
       "      <th>PAY_AMT1</th>\n",
       "      <th>PAY_AMT2</th>\n",
       "      <th>PAY_AMT3</th>\n",
       "      <th>PAY_AMT4</th>\n",
       "      <th>PAY_AMT5</th>\n",
       "      <th>PAY_AMT6</th>\n",
       "      <th>default payment next month</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>20000</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>24</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>-1</td>\n",
       "      <td>-1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>689</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>120000</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>26</td>\n",
       "      <td>-1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>3272</td>\n",
       "      <td>3455</td>\n",
       "      <td>3261</td>\n",
       "      <td>0</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>0</td>\n",
       "      <td>2000</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>90000</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>34</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>14331</td>\n",
       "      <td>14948</td>\n",
       "      <td>15549</td>\n",
       "      <td>1518</td>\n",
       "      <td>1500</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>1000</td>\n",
       "      <td>5000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>50000</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>37</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>28314</td>\n",
       "      <td>28959</td>\n",
       "      <td>29547</td>\n",
       "      <td>2000</td>\n",
       "      <td>2019</td>\n",
       "      <td>1200</td>\n",
       "      <td>1100</td>\n",
       "      <td>1069</td>\n",
       "      <td>1000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>50000</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>57</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>-1</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>20940</td>\n",
       "      <td>19146</td>\n",
       "      <td>19131</td>\n",
       "      <td>2000</td>\n",
       "      <td>36681</td>\n",
       "      <td>10000</td>\n",
       "      <td>9000</td>\n",
       "      <td>689</td>\n",
       "      <td>679</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 25 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   ID  LIMIT_BAL  SEX  EDUCATION  MARRIAGE  AGE  PAY_0  PAY_2  PAY_3  PAY_4  \\\n",
       "0   1      20000    2          2         1   24      2      2     -1     -1   \n",
       "1   2     120000    2          2         2   26     -1      2      0      0   \n",
       "2   3      90000    2          2         2   34      0      0      0      0   \n",
       "3   4      50000    2          2         1   37      0      0      0      0   \n",
       "4   5      50000    1          2         1   57     -1      0     -1      0   \n",
       "\n",
       "              ...              BILL_AMT4  BILL_AMT5  BILL_AMT6  PAY_AMT1  \\\n",
       "0             ...                      0          0          0         0   \n",
       "1             ...                   3272       3455       3261         0   \n",
       "2             ...                  14331      14948      15549      1518   \n",
       "3             ...                  28314      28959      29547      2000   \n",
       "4             ...                  20940      19146      19131      2000   \n",
       "\n",
       "   PAY_AMT2  PAY_AMT3  PAY_AMT4  PAY_AMT5  PAY_AMT6  \\\n",
       "0       689         0         0         0         0   \n",
       "1      1000      1000      1000         0      2000   \n",
       "2      1500      1000      1000      1000      5000   \n",
       "3      2019      1200      1100      1069      1000   \n",
       "4     36681     10000      9000       689       679   \n",
       "\n",
       "   default payment next month  \n",
       "0                           1  \n",
       "1                           1  \n",
       "2                           0  \n",
       "3                           0  \n",
       "4                           0  \n",
       "\n",
       "[5 rows x 25 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([u'LIMIT_BAL', u'SEX', u'EDUCATION', u'MARRIAGE', u'AGE', u'PAY_0',\n",
       "       u'PAY_2', u'PAY_3', u'PAY_4', u'PAY_5', u'PAY_6', u'BILL_AMT1',\n",
       "       u'BILL_AMT2', u'BILL_AMT3', u'BILL_AMT4', u'BILL_AMT5',\n",
       "       u'BILL_AMT6', u'PAY_AMT1', u'PAY_AMT2', u'PAY_AMT3', u'PAY_AMT4',\n",
       "       u'PAY_AMT5', u'PAY_AMT6'], dtype=object)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#get column names as predictors\n",
    "Xnames = data.columns.values[1:-1]\n",
    "Xnames"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "#create train/test mask\n",
    "itrain, itest = train_test_split(xrange(data.shape[0]), train_size=0.6)\n",
    "mask=np.ones(data.shape[0], dtype='int')\n",
    "mask[itrain]=1\n",
    "mask[itest]=0\n",
    "mask = (mask==1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#optimizing cv\n",
    "\n",
    "def cv_optimize(clf, parameters, X, y, n_jobs=1, n_folds=5, score_func=None):\n",
    "    if score_func:\n",
    "        gs = GridSearchCV(clf, param_grid=parameters, cv=n_folds, n_jobs=n_jobs, scoring=score_func)\n",
    "    else:\n",
    "        gs = GridSearchCV(clf, param_grid=parameters, n_jobs=n_jobs, cv=n_folds)\n",
    "    gs.fit(X, y)\n",
    "    print \"BEST\", gs.best_params_, gs.best_score_, gs.grid_scores_\n",
    "    best = gs.best_estimator_\n",
    "    return best\n",
    "\n",
    "# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - #\n",
    "# Important parameters\n",
    "# indf - Input dataframe\n",
    "# featurenames - vector of names of predictors\n",
    "# targetname - name of column you want to predict (e.g. 0 or 1, 'M' or 'F', \n",
    "#              'yes' or 'no')\n",
    "# target1val - particular value you want to have as a 1 in the target\n",
    "# mask - boolean vector indicating test set (~mask is training set)\n",
    "# reuse_split - dictionary that contains traning and testing dataframes \n",
    "#              (we'll use this to test different classifiers on the same \n",
    "#              test-train splits)\n",
    "# score_func - we've used the accuracy as a way of scoring algorithms but \n",
    "#              this can be more general later on\n",
    "# n_folds - Number of folds for cross validation ()\n",
    "# n_jobs - used for parallelization\n",
    "# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - #\n",
    "\n",
    "def do_classify(clf, parameters, indf, featurenames, targetname, target1val, mask=None, reuse_split=None, score_func=None, n_folds=5, n_jobs=1):\n",
    "    subdf=indf[featurenames]\n",
    "    X=subdf.values\n",
    "    y=(indf[targetname].values==target1val)*1\n",
    "    if mask !=None:\n",
    "        print \"using mask\"\n",
    "        Xtrain, Xtest, ytrain, ytest = X[mask], X[~mask], y[mask], y[~mask]\n",
    "    if reuse_split !=None:\n",
    "        print \"using reuse split\"\n",
    "        Xtrain, Xtest, ytrain, ytest = reuse_split['Xtrain'], reuse_split['Xtest'], reuse_split['ytrain'], reuse_split['ytest']\n",
    "    if parameters:\n",
    "        clf = cv_optimize(clf, parameters, Xtrain, ytrain, n_jobs=n_jobs, n_folds=n_folds, score_func=score_func)\n",
    "    clf=clf.fit(Xtrain, ytrain)\n",
    "    training_accuracy = clf.score(Xtrain, ytrain)\n",
    "    test_accuracy = clf.score(Xtest, ytest)\n",
    "    print \"############# based on standard predict ################\"\n",
    "    print \"Accuracy on training data: %0.2f\" % (training_accuracy)\n",
    "    print \"Accuracy on test data:     %0.2f\" % (test_accuracy)\n",
    "    print confusion_matrix(ytest, clf.predict(Xtest))\n",
    "    print \"########################################################\"\n",
    "    return clf, Xtrain, ytrain, Xtest, ytest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Alvin\\Anaconda2\\lib\\site-packages\\ipykernel\\__main__.py:34: FutureWarning: comparison to `None` will result in an elementwise object comparison in the future.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "using mask\n",
      "BEST {'n_estimators': 19} 0.470944423098 [mean: 0.38130, std: 0.01211, params: {'n_estimators': 1}, mean: 0.32306, std: 0.01447, params: {'n_estimators': 2}, mean: 0.41016, std: 0.02220, params: {'n_estimators': 3}, mean: 0.37672, std: 0.02424, params: {'n_estimators': 4}, mean: 0.44319, std: 0.02313, params: {'n_estimators': 5}, mean: 0.40175, std: 0.03101, params: {'n_estimators': 6}, mean: 0.44875, std: 0.02399, params: {'n_estimators': 7}, mean: 0.41374, std: 0.02122, params: {'n_estimators': 8}, mean: 0.45613, std: 0.02829, params: {'n_estimators': 9}, mean: 0.42421, std: 0.02139, params: {'n_estimators': 10}, mean: 0.46063, std: 0.03033, params: {'n_estimators': 11}, mean: 0.43094, std: 0.02507, params: {'n_estimators': 12}, mean: 0.46063, std: 0.01938, params: {'n_estimators': 13}, mean: 0.44327, std: 0.02375, params: {'n_estimators': 14}, mean: 0.46388, std: 0.02598, params: {'n_estimators': 15}, mean: 0.44408, std: 0.02553, params: {'n_estimators': 16}, mean: 0.46570, std: 0.02444, params: {'n_estimators': 17}, mean: 0.45572, std: 0.02882, params: {'n_estimators': 18}, mean: 0.47094, std: 0.02401, params: {'n_estimators': 19}]\n",
      "############# based on standard predict ################\n",
      "Accuracy on training data: 0.99\n",
      "Accuracy on test data:     0.81\n",
      "[[8744  644]\n",
      " [1669  943]]\n",
      "########################################################\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.neighbors import NearestNeighbors\n",
    "\n",
    "#set parameters\n",
    "parameters = {\"n_estimators\": range(1, 20)}\n",
    "clfForest = RandomForestClassifier()\n",
    "\n",
    "clfForest, Xtrain, ytrain, Xtest, ytest = do_classify(clfForest, parameters, \n",
    "                                                       data, Xnames, 'default payment next month', 1, mask=mask, \n",
    "                                                       n_jobs = 4, score_func='f1')   "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![hwimages](./images/confusionmatrix.png)\n",
    "When building machine learning models, there will always be inaccuracies.  In some cases, however, the kind of inaccuracy matters.  For example, if you are trying to predict if a patient has cancer, the model will need to be more conservative and should lean towards classifying patients as having cancer versus not having it -- this means having more False Positives than False Negatives).   In this case, do false positives and false negatives matter?  Credit companies would want to charge customers with a higher probability of default a higher premium.  This means that evaluating the probability would be more useful than the classification itself.  Let's do that."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#optimizing cv\n",
    "\n",
    "def cv_optimize(clf, parameters, X, y, n_jobs=1, n_folds=5, score_func=None):\n",
    "    if score_func:\n",
    "        gs = GridSearchCV(clf, param_grid=parameters, cv=n_folds, n_jobs=n_jobs, scoring=score_func)\n",
    "    else:\n",
    "        gs = GridSearchCV(clf, param_grid=parameters, n_jobs=n_jobs, cv=n_folds)\n",
    "    gs.fit(X, y)\n",
    "    print \"BEST\", gs.best_params_, gs.best_score_, gs.grid_scores_\n",
    "    best = gs.best_estimator_\n",
    "    return best\n",
    "\n",
    "# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - #\n",
    "# Important parameters\n",
    "# indf - Input dataframe\n",
    "# featurenames - vector of names of predictors\n",
    "# targetname - name of column you want to predict (e.g. 0 or 1, 'M' or 'F', \n",
    "#              'yes' or 'no')\n",
    "# target1val - particular value you want to have as a 1 in the target\n",
    "# mask - boolean vector indicating test set (~mask is training set)\n",
    "# reuse_split - dictionary that contains traning and testing dataframes \n",
    "#              (we'll use this to test different classifiers on the same \n",
    "#              test-train splits)\n",
    "# score_func - we've used the accuracy as a way of scoring algorithms but \n",
    "#              this can be more general later on\n",
    "# n_folds - Number of folds for cross validation ()\n",
    "# n_jobs - used for parallelization\n",
    "# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - #\n",
    "\n",
    "def do_proba(clf, parameters, indf, featurenames, targetname, target1val, mask=None, reuse_split=None, score_func=None, n_folds=5, n_jobs=1):\n",
    "    subdf=indf[featurenames]\n",
    "    X=subdf.values\n",
    "    y=(indf[targetname].values==target1val)*1\n",
    "    if mask !=None:\n",
    "        print \"using mask\"\n",
    "        Xtrain, Xtest, ytrain, ytest = X[mask], X[~mask], y[mask], y[~mask]\n",
    "    if reuse_split !=None:\n",
    "        print \"using reuse split\"\n",
    "        Xtrain, Xtest, ytrain, ytest = reuse_split['Xtrain'], reuse_split['Xtest'], reuse_split['ytrain'], reuse_split['ytest']\n",
    "    if parameters:\n",
    "        clf = cv_optimize(clf, parameters, Xtrain, ytrain, n_jobs=n_jobs, n_folds=n_folds, score_func=score_func)\n",
    "    clf=clf.fit(Xtrain, ytrain)\n",
    "    predictions = clf.predict_proba(Xtest)\n",
    "    return predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Alvin\\Anaconda2\\lib\\site-packages\\ipykernel\\__main__.py:34: FutureWarning: comparison to `None` will result in an elementwise object comparison in the future.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "using mask\n",
      "BEST {'n_estimators': 19} 0.471532507346 [mean: 0.38933, std: 0.01431, params: {'n_estimators': 1}, mean: 0.32692, std: 0.02516, params: {'n_estimators': 2}, mean: 0.42392, std: 0.02138, params: {'n_estimators': 3}, mean: 0.37582, std: 0.02486, params: {'n_estimators': 4}, mean: 0.44035, std: 0.01272, params: {'n_estimators': 5}, mean: 0.40700, std: 0.03318, params: {'n_estimators': 6}, mean: 0.44973, std: 0.02957, params: {'n_estimators': 7}, mean: 0.42120, std: 0.01132, params: {'n_estimators': 8}, mean: 0.45755, std: 0.02730, params: {'n_estimators': 9}, mean: 0.42906, std: 0.02361, params: {'n_estimators': 10}, mean: 0.45702, std: 0.02641, params: {'n_estimators': 11}, mean: 0.43463, std: 0.02705, params: {'n_estimators': 12}, mean: 0.45439, std: 0.02955, params: {'n_estimators': 13}, mean: 0.44501, std: 0.02241, params: {'n_estimators': 14}, mean: 0.45980, std: 0.02702, params: {'n_estimators': 15}, mean: 0.44937, std: 0.02689, params: {'n_estimators': 16}, mean: 0.46431, std: 0.02813, params: {'n_estimators': 17}, mean: 0.45513, std: 0.03308, params: {'n_estimators': 18}, mean: 0.47153, std: 0.02870, params: {'n_estimators': 19}]\n"
     ]
    }
   ],
   "source": [
    "probs = do_proba(clfForest, parameters, data, Xnames, 'default payment next month', 1, mask=mask, n_jobs = 4, score_func='f1')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Relative Importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbYAAAEZCAYAAAD7QwBLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCw19AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0xaf32630>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "importance_list = clfForest.feature_importances_\n",
    "name_list = data.columns\n",
    "importance_list, name_list = zip(*sorted(zip(importance_list, name_list)))\n",
    "plt.barh(range(len(name_list)),importance_list,align='center')\n",
    "plt.yticks(range(len(name_list)),name_list)\n",
    "plt.xlabel('Relative Importance in the Random Forest')\n",
    "plt.ylabel('Features')\n",
    "plt.title('Relative importance of Each Feature')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.42105263,  0.57894737],\n",
       "       [ 0.73684211,  0.26315789],\n",
       "       [ 0.84210526,  0.15789474],\n",
       "       ..., \n",
       "       [ 0.47368421,  0.52631579],\n",
       "       [ 1.        ,  0.        ],\n",
       "       [ 0.89473684,  0.10526316]])"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "probs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [Root]",
   "language": "python",
   "name": "Python [Root]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}