Credit Card Fraud Final.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Credit Card Fraud Detection (Imbalanced dataset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Populating the interactive namespace from numpy and matplotlib\n"
     ]
    }
   ],
   "source": [
    "%pylab inline\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.metrics import accuracy_score, roc_auc_score, confusion_matrix, classification_report\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import cross_val_predict\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Read the data file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Read the data file\n",
    "cc = pd.read_csv(\"E:/SCU Business Analytics/Machine Learning/Assignment/DSTMAA_data/creditcard.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "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>Time</th>\n",
       "      <th>V1</th>\n",
       "      <th>V2</th>\n",
       "      <th>V3</th>\n",
       "      <th>V4</th>\n",
       "      <th>V5</th>\n",
       "      <th>V6</th>\n",
       "      <th>V7</th>\n",
       "      <th>V8</th>\n",
       "      <th>V9</th>\n",
       "      <th>...</th>\n",
       "      <th>V21</th>\n",
       "      <th>V22</th>\n",
       "      <th>V23</th>\n",
       "      <th>V24</th>\n",
       "      <th>V25</th>\n",
       "      <th>V26</th>\n",
       "      <th>V27</th>\n",
       "      <th>V28</th>\n",
       "      <th>Amount</th>\n",
       "      <th>Class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>-1.359807</td>\n",
       "      <td>-0.072781</td>\n",
       "      <td>2.536347</td>\n",
       "      <td>1.378155</td>\n",
       "      <td>-0.338321</td>\n",
       "      <td>0.462388</td>\n",
       "      <td>0.239599</td>\n",
       "      <td>0.098698</td>\n",
       "      <td>0.363787</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.018307</td>\n",
       "      <td>0.277838</td>\n",
       "      <td>-0.110474</td>\n",
       "      <td>0.066928</td>\n",
       "      <td>0.128539</td>\n",
       "      <td>-0.189115</td>\n",
       "      <td>0.133558</td>\n",
       "      <td>-0.021053</td>\n",
       "      <td>149.62</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.191857</td>\n",
       "      <td>0.266151</td>\n",
       "      <td>0.166480</td>\n",
       "      <td>0.448154</td>\n",
       "      <td>0.060018</td>\n",
       "      <td>-0.082361</td>\n",
       "      <td>-0.078803</td>\n",
       "      <td>0.085102</td>\n",
       "      <td>-0.255425</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.225775</td>\n",
       "      <td>-0.638672</td>\n",
       "      <td>0.101288</td>\n",
       "      <td>-0.339846</td>\n",
       "      <td>0.167170</td>\n",
       "      <td>0.125895</td>\n",
       "      <td>-0.008983</td>\n",
       "      <td>0.014724</td>\n",
       "      <td>2.69</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>-1.358354</td>\n",
       "      <td>-1.340163</td>\n",
       "      <td>1.773209</td>\n",
       "      <td>0.379780</td>\n",
       "      <td>-0.503198</td>\n",
       "      <td>1.800499</td>\n",
       "      <td>0.791461</td>\n",
       "      <td>0.247676</td>\n",
       "      <td>-1.514654</td>\n",
       "      <td>...</td>\n",
       "      <td>0.247998</td>\n",
       "      <td>0.771679</td>\n",
       "      <td>0.909412</td>\n",
       "      <td>-0.689281</td>\n",
       "      <td>-0.327642</td>\n",
       "      <td>-0.139097</td>\n",
       "      <td>-0.055353</td>\n",
       "      <td>-0.059752</td>\n",
       "      <td>378.66</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>-0.966272</td>\n",
       "      <td>-0.185226</td>\n",
       "      <td>1.792993</td>\n",
       "      <td>-0.863291</td>\n",
       "      <td>-0.010309</td>\n",
       "      <td>1.247203</td>\n",
       "      <td>0.237609</td>\n",
       "      <td>0.377436</td>\n",
       "      <td>-1.387024</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.108300</td>\n",
       "      <td>0.005274</td>\n",
       "      <td>-0.190321</td>\n",
       "      <td>-1.175575</td>\n",
       "      <td>0.647376</td>\n",
       "      <td>-0.221929</td>\n",
       "      <td>0.062723</td>\n",
       "      <td>0.061458</td>\n",
       "      <td>123.50</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2.0</td>\n",
       "      <td>-1.158233</td>\n",
       "      <td>0.877737</td>\n",
       "      <td>1.548718</td>\n",
       "      <td>0.403034</td>\n",
       "      <td>-0.407193</td>\n",
       "      <td>0.095921</td>\n",
       "      <td>0.592941</td>\n",
       "      <td>-0.270533</td>\n",
       "      <td>0.817739</td>\n",
       "      <td>...</td>\n",
       "      <td>-0.009431</td>\n",
       "      <td>0.798278</td>\n",
       "      <td>-0.137458</td>\n",
       "      <td>0.141267</td>\n",
       "      <td>-0.206010</td>\n",
       "      <td>0.502292</td>\n",
       "      <td>0.219422</td>\n",
       "      <td>0.215153</td>\n",
       "      <td>69.99</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Time        V1        V2        V3        V4        V5        V6        V7  \\\n",
       "0   0.0 -1.359807 -0.072781  2.536347  1.378155 -0.338321  0.462388  0.239599   \n",
       "1   0.0  1.191857  0.266151  0.166480  0.448154  0.060018 -0.082361 -0.078803   \n",
       "2   1.0 -1.358354 -1.340163  1.773209  0.379780 -0.503198  1.800499  0.791461   \n",
       "3   1.0 -0.966272 -0.185226  1.792993 -0.863291 -0.010309  1.247203  0.237609   \n",
       "4   2.0 -1.158233  0.877737  1.548718  0.403034 -0.407193  0.095921  0.592941   \n",
       "\n",
       "         V8        V9  ...         V21       V22       V23       V24  \\\n",
       "0  0.098698  0.363787  ...   -0.018307  0.277838 -0.110474  0.066928   \n",
       "1  0.085102 -0.255425  ...   -0.225775 -0.638672  0.101288 -0.339846   \n",
       "2  0.247676 -1.514654  ...    0.247998  0.771679  0.909412 -0.689281   \n",
       "3  0.377436 -1.387024  ...   -0.108300  0.005274 -0.190321 -1.175575   \n",
       "4 -0.270533  0.817739  ...   -0.009431  0.798278 -0.137458  0.141267   \n",
       "\n",
       "        V25       V26       V27       V28  Amount  Class  \n",
       "0  0.128539 -0.189115  0.133558 -0.021053  149.62      0  \n",
       "1  0.167170  0.125895 -0.008983  0.014724    2.69      0  \n",
       "2 -0.327642 -0.139097 -0.055353 -0.059752  378.66      0  \n",
       "3  0.647376 -0.221929  0.062723  0.061458  123.50      0  \n",
       "4 -0.206010  0.502292  0.219422  0.215153   69.99      0  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cc.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Data Exploration"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "28 out of 31 columns are PCA generated. We do not have original features and more background information about the data (V1,V2,....V28).  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of observations: 284807\n"
     ]
    }
   ],
   "source": [
    "print('Total number of observations:',len(cc.index))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Percentage of fradulent transactions in data.: 0.1727485630620034\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1e2000b88d0>"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print('Percentage of fradulent transactions in data.:',cc.Class.mean()*100)\n",
    "sns.countplot(cc.Class)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It is clear from the above  distribution that number of fradulent transactions are very low (only .17%) as compared to non-fradulent transaction (99.83%). Directly running any classifier will yield higher accuracy score as one of the class is highly represented in the data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Time - Data is collected for 2 days. During the day time, number of transactions are higher than night. Will drop this variable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1e2000efe48>"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(cc.Time)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Amount - There are very few transactions of higher values. Lower value transactions dominates the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1e200153518>"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(cc.Amount,kde=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Normalizing the variables (Amount)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "cc['amt_normalize'] = ((cc.Amount) - (cc.Amount.mean()))/(cc.Amount.std())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b><font color = 'blue'>A. Directly Running logistic regression to classify fradulent and non-fradulent transactions. \n",
    "<br><b><p> <font color='red'>This approach is not correct. It's just to showcase what is wrong."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>A-1. Creating training and test data  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "cc1 = cc.copy()\n",
    "y = cc.Class\n",
    "X = cc.drop(columns =['Class','Amount','Time']) #dropping time too\n",
    "#  Train and Test datasets\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "#saving original data of X and y. This will be used when trying out diffrent methods.\n",
    "X_org = X\n",
    "y_org = y"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>A-2. Fit logistic regression classifier to train data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "# instantiate a logistic regression model, and fit with X_train and y_train\n",
    "model = LogisticRegression()\n",
    "model = model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>A-3. Accuracy, Confusion Matrix & is there Overfitting?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy on the training set: 0.9992191594172518\n",
      "accuracy on the testing set: 0.9991594582229457\n"
     ]
    }
   ],
   "source": [
    "# check the accuracy on the training set\n",
    "print(\"accuracy on the training set:\",model.score(X_train, y_train))\n",
    "\n",
    "# check the accuracy on the test dataset\n",
    "print(\"accuracy on the testing set:\",model.score(X_test, y_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix on train data\n",
      "[[190463    122]\n",
      " [    27    208]]\n",
      "Confusion Matrix on test data\n",
      "[[93811    65]\n",
      " [   14    97]]\n"
     ]
    }
   ],
   "source": [
    "# Predict class labels for the train set\n",
    "predicted = model.predict(X_train)\n",
    "# Confusion Matrix for train data\n",
    "print('Confusion Matrix on train data')\n",
    "print(confusion_matrix(predicted, y_train))\n",
    "\n",
    "# Predict class labels for the test dataset\n",
    "predicted_test = model.predict(X_test)\n",
    "# Confusion Matrix for test data\n",
    "print('Confusion Matrix on test data')\n",
    "print(confusion_matrix(predicted_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There is <b>no overfitting</b> because the model is predicting the 'test' dataset almost similar to train dataset.\n",
    "<br><br>\n",
    "Since,our dataset is highly imbalanced, the accuracy is only reflecting the underlying class distribution which in this case is majority non-fradulent cases. Therefore, we need to compute other metrices like precision, recall and F1 score using confusion matrix to detect overfitting."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>A-4. Precision, Recall, F1-Scores and AUC."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     93876\n",
      "           1       0.60      0.87      0.71       111\n",
      "\n",
      "   micro avg       1.00      1.00      1.00     93987\n",
      "   macro avg       0.80      0.94      0.86     93987\n",
      "weighted avg       1.00      1.00      1.00     93987\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(predicted_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We observed that our model is predicting non-fraulent transactions very well. However, they are <b> not predicting fradulent transactions</b> that well. For fradulent transactions, precision is 60%, recall is 87% and F1-score of 71% only."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>So, which metric is most important?</b>\n",
    "<ul><li>Precision is an important metric - We want our model to be precise. If our model is not predicting correctly then banks would block the genuine transactions which lead to customer dissatisfaction.\n",
    "    <li>Recall is also an important metric - If recall is low than our model is not predicting fradulent transactions well. Bank will have to incur losses as these transactions are not getting detected correctly.\n",
    "    <li>Therefore, we are looking at F1 score which is harmonic mean of precision and recall scores."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area under curve (AUC):  0.96833762619534\n"
     ]
    }
   ],
   "source": [
    "#ROC, AUC\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "y_score = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = roc_curve(y_test, y_score)\n",
    "\n",
    "title('ROC curve')\n",
    "xlabel('FPR (Precision)')\n",
    "ylabel('TPR (Recall)')\n",
    "\n",
    "plot(fpr,tpr)\n",
    "plot((0,1), ls='dashed',color='black')\n",
    "plt.show()\n",
    "print('Area under curve (AUC): ', auc(fpr,tpr))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Though AUC is high, we are not satisfied with the our results as it is not capturing fraudulent transactions that well. Our model is predicting non-fradulent transaction better than fradulent transactions. We also know that this is happening because of highly imbalanced dataset. So what should we do?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Following are few techniques I have tried to cater the issue of imbalanced dataset.\n",
    "<ol><li> Cross-Validation\n",
    "    <li> Oversample minority class (i.e. fradulent transactions)\n",
    "    <li> Ensemble modelling (Random Forrest)\n",
    "    <li> Bagging\n",
    "    <li> Boosting (XGBoost, ADABoost)\n",
    "    <li> Changing the Threshold"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b><font color = 'blue'>1. Cross-Validation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On Train data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix on train data\n",
      "[[190463     27]\n",
      " [   126    204]]\n",
      "accuracy on the training set: 0.9991981972539566\n"
     ]
    }
   ],
   "source": [
    "scores = cross_val_score(model, X_train, y_train, cv=10)\n",
    "y_pred = cross_val_predict(model, X_train, y_train, cv=10)\n",
    "conf_mat = confusion_matrix(y_train, y_pred)\n",
    "print('Confusion Matrix on train data')\n",
    "print(conf_mat)\n",
    "# Accuracy on train data:\n",
    "print(\"accuracy on the training set:\",scores.mean())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On Test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix on test data\n",
      "[[93808    17]\n",
      " [   67    95]]\n",
      "accuracy on the training set: 0.9991062584641602\n"
     ]
    }
   ],
   "source": [
    "scores_test = cross_val_score(model, X_test, y_test, cv=10)\n",
    "y_pred_test = cross_val_predict(model, X_test, y_test, cv=10)\n",
    "conf_mat_test = confusion_matrix(y_test, y_pred_test)\n",
    "print('Confusion Matrix on test data')\n",
    "print(conf_mat_test)\n",
    "\n",
    "#Accuracy on Test Data:\n",
    "print(\"accuracy on the training set:\",scores_test.mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     93875\n",
      "           1       0.59      0.85      0.69       112\n",
      "\n",
      "   micro avg       1.00      1.00      1.00     93987\n",
      "   macro avg       0.79      0.92      0.85     93987\n",
      "weighted avg       1.00      1.00      1.00     93987\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# cross validation- classification report on test data:\n",
    "print(classification_report(y_pred_test, y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>There is not much improvement using cross-validation. For fradulent transactions, precision is 59%, recall is 85% and F1-score of 69% only."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b><font color = 'blue'>2. Oversampling minority class (i.e. fradulent transactions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [],
   "source": [
    "from imblearn.over_sampling import SMOTE\n",
    "sm = SMOTE(random_state=42)\n",
    "X= X_org\n",
    "y= y_org\n",
    "X_res, y_res = sm.fit_resample(X, y)\n",
    "\n",
    "#  Train and Test datasets\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_res, y_res, test_size=0.33, random_state=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Fitting the model on train data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [],
   "source": [
    "# instantiate a logistic regression model, and fit with X_train and y_train\n",
    "model = LogisticRegression()\n",
    "model = model.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy on the testing set: 0.9464955661664393\n",
      "[[91344  7691]\n",
      " [ 2349 86264]]\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.97      0.92      0.95     99035\n",
      "           1       0.92      0.97      0.95     88613\n",
      "\n",
      "   micro avg       0.95      0.95      0.95    187648\n",
      "   macro avg       0.95      0.95      0.95    187648\n",
      "weighted avg       0.95      0.95      0.95    187648\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area under curve (AUC):  0.9890538533977183\n"
     ]
    }
   ],
   "source": [
    "# check the accuracy on the test dataset\n",
    "print(\"accuracy on the testing set:\",model.score(X_test, y_test))\n",
    "\n",
    "# Predict class labels for the test set\n",
    "predicted = model.predict(X_test)\n",
    "\n",
    "# Confusion Matrix for train data\n",
    "print(confusion_matrix(predicted, y_test))\n",
    "\n",
    "print(classification_report(predicted, y_test))\n",
    "\n",
    "#ROC, AUC\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "y_score = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = roc_curve(y_test, y_score)\n",
    "\n",
    "title('ROC curve')\n",
    "xlabel('FPR (Precision)')\n",
    "ylabel('TPR (Recall)')\n",
    "\n",
    "plot(fpr,tpr)\n",
    "plot((0,1), ls='dashed',color='black')\n",
    "plt.show()\n",
    "print('Area under curve (AUC): ', auc(fpr,tpr))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>AUC is high (98.9%). Along with it our F1-score, precison and recall scores are also high (95%, 92% and 97% respectively).\n",
    "\n",
    "With this approach, we are able to correctly predict 92% of fradulent cases correctly. After oversampling the minority class, our detection rate improved significantly."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b><font color = 'blue'> 3. Ensemble modelling (Random Forrest)</b>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [],
   "source": [
    "#preparing original data\n",
    "X= X_org\n",
    "y= y_org\n",
    "\n",
    "#  Train and Test datasets\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import ensemble\n",
    "rf = ensemble.RandomForestClassifier()\n",
    "model_rf = rf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AUC = 0.9287169112446668\n",
      "Accuracy = 0.9994786513028397\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     93825\n",
      "           1       0.93      0.75      0.83       162\n",
      "\n",
      "   micro avg       1.00      1.00      1.00     93987\n",
      "   macro avg       0.97      0.88      0.92     93987\n",
      "weighted avg       1.00      1.00      1.00     93987\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area under curve (AUC):  0.9287169112446668\n"
     ]
    }
   ],
   "source": [
    "predicted_rf = model_rf.predict(X_test)\n",
    "probs_rf = model_rf.predict_proba(X_test)\n",
    "print('AUC =', roc_auc_score(y_test, probs_rf[:, 1]))\n",
    "print('Accuracy =', accuracy_score(y_test, predicted_rf))\n",
    "print(classification_report(y_test, predicted_rf))\n",
    "\n",
    "#ROC, AUC\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "y_score = model_rf.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = roc_curve(y_test, y_score)\n",
    "\n",
    "title('ROC curve')\n",
    "xlabel('FPR (Precision)')\n",
    "ylabel('TPR (Recall)')\n",
    "\n",
    "plot(fpr,tpr)\n",
    "plot((0,1), ls='dashed',color='black')\n",
    "plt.show()\n",
    "print('Area under curve (AUC): ', auc(fpr,tpr))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>With Random forrest, our AUC has dropped to 92.2% from 96.8% and recall has also seen a dropped to 75% from 85%. However, precision has been increased from 60% to 94%. If we are only concerned about precision then ensemble method like Random forrest gives better results."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b><font color = 'blue'>4. Bagging"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [],
   "source": [
    "#preparing original data\n",
    "X= X_org\n",
    "y= y_org\n",
    "\n",
    "#  Train and Test datasets\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     93879\n",
      "           1       0.59      0.88      0.70       108\n",
      "\n",
      "   micro avg       1.00      1.00      1.00     93987\n",
      "   macro avg       0.79      0.94      0.85     93987\n",
      "weighted avg       1.00      1.00      1.00     93987\n",
      "\n",
      "Accuracy = 0.9991488184536159\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area under curve (AUC):  0.9680629488179004\n"
     ]
    }
   ],
   "source": [
    "from sklearn import model_selection\n",
    "from sklearn.ensemble import BaggingClassifier\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "\n",
    "seed = 7\n",
    "kfold = model_selection.KFold(n_splits=10, random_state=seed)\n",
    "num_trees = 100\n",
    "model = BaggingClassifier(LogisticRegression())\n",
    "model.fit(X_train, y_train)\n",
    "\n",
    "#test set\n",
    "predicted = model.predict(X_test)\n",
    "print(classification_report(predicted, y_test))\n",
    "print('Accuracy =', accuracy_score(y_test, predicted))\n",
    "\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "y_score = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = roc_curve(y_test, y_score)\n",
    "\n",
    "title('ROC curve')\n",
    "xlabel('FPR (Precision)')\n",
    "ylabel('TPR (Recall)')\n",
    "\n",
    "plot(fpr,tpr)\n",
    "plot((0,1), ls='dashed',color='black')\n",
    "plt.show()\n",
    "print('Area under curve (AUC): ', auc(fpr,tpr))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Bagging does not play much role in imporvements of AUC, recall or precision scoes. There are marginal improvements."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b><font color = 'blue'>5. Boosting (XGBoost, ADABoost)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<B>5.1 XG Boost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "#preparing original data\n",
    "X= X_org\n",
    "y= y_org\n",
    "\n",
    "#  Train and Test datasets\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy on the testing set: 0.999521210380159\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       1.00      1.00      1.00     93848\n",
      "           1       0.79      0.92      0.85       139\n",
      "\n",
      "   micro avg       1.00      1.00      1.00     93987\n",
      "   macro avg       0.90      0.96      0.93     93987\n",
      "weighted avg       1.00      1.00      1.00     93987\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area under curve (AUC):  0.9828346376396825\n"
     ]
    }
   ],
   "source": [
    "from xgboost import XGBClassifier\n",
    "model2 = XGBClassifier(objective='reg:logistic')\n",
    "model2.fit(X_train, y_train)\n",
    "\n",
    "y_pred1 = model2.predict(X_test)\n",
    "# check the accuracy on the test dataset\n",
    "print(\"accuracy on the testing set:\",model2.score(X_test, y_test))\n",
    "print(classification_report(y_pred1, y_test))\n",
    "\n",
    "#ROC, AUC\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "y_score = model2.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = roc_curve(y_test, y_score)\n",
    "\n",
    "title('ROC curve')\n",
    "xlabel('FPR (Precision)')\n",
    "ylabel('TPR (Recall)')\n",
    "\n",
    "plot(fpr,tpr)\n",
    "plot((0,1), ls='dashed',color='black')\n",
    "plt.show()\n",
    "print('Area under curve (AUC): ', auc(fpr,tpr))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Xgboost increased the performance of f1 score significantly to 85% from 71%(Refer section A-4 - original model) and AUC increased from 96.8% to 98.28%."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>5.2 ADABOOST"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [],
   "source": [
    "#preparing original data\n",
    "X= X_org\n",
    "y= y_org\n",
    "\n",
    "#  Train and Test datasets\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "boost = AdaBoostClassifier()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[93804    21]\n",
      " [   56   106]]\n",
      "accuracy on the testing set: 0.9991807377616053\n",
      "Accuracy = 0.9991807377616053\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0  0.99940337 0.99977618 0.99958974     93825\n",
      "           1  0.83464567 0.65432099 0.73356401       162\n",
      "\n",
      "   micro avg  0.99918074 0.99918074 0.99918074     93987\n",
      "   macro avg  0.91702452 0.82704858 0.86657688     93987\n",
      "weighted avg  0.99911938 0.99918074 0.99913120     93987\n",
      "\n",
      "[0.44843479 0.44348364 0.44381987 ... 0.45341238 0.26069472 0.45121132]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area under curve (AUC):  0.9698149957400335\n"
     ]
    }
   ],
   "source": [
    "boost.fit(X_train, y_train)\n",
    "y_pred_boost = boost.predict(X_test)\n",
    "print(confusion_matrix( y_test,y_pred_boost))\n",
    "print(\"accuracy on the testing set:\",boost.score(X_test, y_test))\n",
    "print('Accuracy =', accuracy_score(y_test, y_pred_boost))\n",
    "print(classification_report(y_test,y_pred_boost,digits=8))\n",
    "\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "y_score = boost.predict_proba(X_test)[:,1]\n",
    "print(y_score)\n",
    "fpr, tpr, _ = roc_curve(y_test, y_score)\n",
    "\n",
    "title('ROC curve')\n",
    "xlabel('FPR ')\n",
    "ylabel('TPR (Recall)')\n",
    "\n",
    "plot(fpr,tpr)\n",
    "plot((0,1), ls='dashed',color='black')\n",
    "plt.show()\n",
    "print('Area under curve (AUC): ', auc(fpr,tpr))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Adaboost does not play much role in imporvements of AUC, recall and F1-scores. There are marginal improvements. However,  precision scores have been improved from 60% to 83%."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b><font color = 'blue'>6. Changing the Threshold"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[93799    34]\n",
      " [   26   128]]\n",
      "Accuracy = 0.9993616138402119\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "       False       1.00      1.00      1.00     93833\n",
      "        True       0.79      0.83      0.81       154\n",
      "\n",
      "   micro avg       1.00      1.00      1.00     93987\n",
      "   macro avg       0.89      0.92      0.90     93987\n",
      "weighted avg       1.00      1.00      1.00     93987\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Area under curve (AUC):  0.96833762619534\n"
     ]
    }
   ],
   "source": [
    "X= X_org\n",
    "y= y_org\n",
    "#  Train and Test datasets\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=0)\n",
    "\n",
    "# instantiate a logistic regression model, and fit with X_train and y_train\n",
    "model = LogisticRegression()\n",
    "model = model.fit(X_train, y_train)\n",
    "\n",
    "# check the accuracy on the training set\n",
    "model.score(X_train, y_train)\n",
    "\n",
    "# Predict class labels for the test set\n",
    "predicted = (model.predict_proba(X_test)[:,1] >= 0.1).astype(bool)\n",
    "\n",
    "# Confusion Matrix for train data\n",
    "print(confusion_matrix(predicted, y_test))\n",
    "print('Accuracy =', accuracy_score(y_test, predicted))\n",
    "\n",
    "print(classification_report(predicted, y_test))\n",
    "\n",
    "#ROC, AUC\n",
    "from sklearn.metrics import roc_curve, auc\n",
    "y_score = model.predict_proba(X_test)[:,1]\n",
    "fpr, tpr, _ = roc_curve(y_test, y_score)\n",
    "\n",
    "title('ROC curve')\n",
    "xlabel('FPR (Precision)')\n",
    "ylabel('TPR (Recall)')\n",
    "\n",
    "plot(fpr,tpr)\n",
    "plot((0,1), ls='dashed',color='black')\n",
    "plt.show()\n",
    "print('Area under curve (AUC): ', auc(fpr,tpr))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>After changing the threshold to 10%, our F1-score for identifying fraudulent transactions has improved to 81% from 71%(Refer section A-4 - original model) though AUC remains the same."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b><font color='blue'>CONCLUSION"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Scores summary of all the models are as follows:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<table style=\"width:100%\">\n",
    "  <tr>\n",
    "    <th>Model</th>\n",
    "    <th>Accuracy</th>\n",
    "    <th>AUC</th> \n",
    "    <th>Precision</th>\n",
    "    <th>Recall</th>\n",
    "    <th>F1-Score</th>\n",
    "  </tr>\n",
    "  <tr>\n",
    "    <td>Original Model</td>\n",
    "    <td>0.99</td> \n",
    "    <td>0.96</td>\n",
    "      <td>0.6</td>\n",
    "      <td>0.87</td>\n",
    "      <td>0.71</td>\n",
    "  </tr>\n",
    "    \n",
    "  <tr>\n",
    "    <td>Cross Validation</td>\n",
    "    <td>0.99</td> \n",
    "    <td></td>\n",
    "      <td>0.59</td>\n",
    "      <td>0.85</td>\n",
    "      <td>0.69</td>\n",
    "  </tr>\n",
    "  \n",
    "  <tr bgcolor=\"#32CD32\">\n",
    "    <td>Oversampling (SMOTE)</td>\n",
    "    <td>0.946</td> \n",
    "    <td>0.989</td>\n",
    "      <td>0.92</td>\n",
    "      <td>0.97</td>\n",
    "      <td>0.95</td>\n",
    "  </tr>\n",
    "  \n",
    "   <tr>\n",
    "    <td>Ensemble (Random Forrest)</td>\n",
    "    <td>0.99</td> \n",
    "    <td>0.928</td>\n",
    "      <td>0.93</td>\n",
    "      <td>0.75</td>\n",
    "      <td>0.83</td>\n",
    "  </tr>\n",
    "  \n",
    "   <tr>\n",
    "    <td>Bagging (Decision Trees)</td>\n",
    "    <td>0.99</td> \n",
    "    <td>0.97</td>\n",
    "      <td>0.61</td>\n",
    "      <td>0.88</td>\n",
    "      <td>0.72</td>\n",
    "  </tr>\n",
    "  \n",
    "  <tr>\n",
    "    <td>Boosting (XGBOOST)</td>\n",
    "    <td>0.99</td> \n",
    "    <td>0.98</td>\n",
    "      <td>0.79</td>\n",
    "      <td>0.92</td>\n",
    "      <td>0.85</td>\n",
    "  </tr>\n",
    "  \n",
    "  <tr>\n",
    "    <td>Boosting (ADABOOST)</td>\n",
    "    <td>0.99</td> \n",
    "    <td>0.969</td>\n",
    "      <td>0.83</td>\n",
    "      <td>0.65</td>\n",
    "      <td>0.73</td>\n",
    "  </tr>\n",
    "  \n",
    "  <tr>\n",
    "    <td>Threshold change to 10%</td>\n",
    "    <td>0.99</td> \n",
    "    <td>0.968</td>\n",
    "      <td>0.79</td>\n",
    "      <td>0.83</td>\n",
    "      <td>0.81</td>\n",
    "  </tr>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Out of all the methods we tried, we found that oversampling(SMOTE) method is best suited for our analysis. Significant high F1, Precision and recall scores are observed at a cost of small decline in the accuracy score.<br>\n",
    "\n",
    "Thus, with use of oversampling(SMOTE) model, banks can better detect fradulent transactions."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "--------------------------------------- End of the document ---------------------------------------"
   ]
  }
 ],
 "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}