PredictiveCreditRisk.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "\n",
    "## **Predictive Credit Risk Model**\n",
    "__author__ : Jebin Jervis \n",
    "\n",
    "This notebook demonstrates a **Machine Learning approach** to predict credit risk using the UCI Default of Credit Card Clients dataset. It includes data preprocessing, model training, and evaluation using a Random Forest Classifier.\n",
    "\n",
    "---\n",
    "\n",
    "### **1. Dataset Overview**\n",
    "\n",
    "The dataset includes the following key features:\n",
    "- **ID**: Unique client ID\n",
    "- **LIMIT_BAL**: Credit limit\n",
    "- **AGE**: Client's age\n",
    "- **SEX, EDUCATION, MARRIAGE**: Categorical demographic variables\n",
    "- **PAY_0 to PAY_6**: Payment status in previous months\n",
    "- **BILL_AMT1 to BILL_AMT6**: Bill statements for the past 6 months\n",
    "- **PAY_AMT1 to PAY_AMT6**: Amount paid in previous months\n",
    "- **default**: Target variable (1 = Default, 0 = No Default)\n",
    "\n",
    "---\n",
    "\n",
    "### **2. Data Preprocessing**\n",
    "\n",
    "Key preprocessing steps include:\n",
    "1. **Handling Missing Values**: Replacing NaN with column means.\n",
    "2. **Standardizing Numeric Columns**: Scaling features like `LIMIT_BAL`, `AGE`, and `BILL_AMT` columns using `StandardScaler`.\n",
    "3. **Encoding Categorical Variables**: Label Encoding applied to `SEX`, `EDUCATION`, `MARRIAGE`, and payment status columns (`PAY_0` to `PAY_6`).\n",
    "\n",
    "---\n",
    "\n",
    "### **3. Class Imbalance Check**\n",
    "\n",
    "- The dataset has an **equal number of \"Default\" and \"No Default\" cases**.\n",
    "- This confirms **no significant class imbalance**, so no resampling techniques like SMOTE are applied.\n",
    "\n",
    "---\n",
    "\n",
    "### **4. Model Training**\n",
    "\n",
    "- **Model**: Random Forest Classifier\n",
    "- **Hyperparameter Tuning**: GridSearchCV is used for:\n",
    "  - `n_estimators`: Number of trees in the forest\n",
    "  - `max_depth`: Maximum tree depth\n",
    "  - `min_samples_split`: Minimum samples to split an internal node\n",
    "  - `max_features`: Number of features to consider at each split\n",
    "\n",
    "---\n",
    "\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 3 folds for each of 24 candidates, totalling 72 fits\n",
      "Best Parameters: {'max_depth': None, 'max_features': 'sqrt', 'min_samples_split': 2, 'n_estimators': 200}\n",
      "Accuracy: 0.8544832013695699\n",
      "ROC AUC Score: 0.9244718659844523\n",
      "\n",
      "Classification Report:\n",
      "               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.85      0.86      0.86      4673\n",
      "           1       0.86      0.85      0.85      4673\n",
      "\n",
      "    accuracy                           0.85      9346\n",
      "   macro avg       0.85      0.85      0.85      9346\n",
      "weighted avg       0.85      0.85      0.85      9346\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Import libraries\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split, GridSearchCV\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import accuracy_score, roc_auc_score, classification_report, confusion_matrix\n",
    "from sklearn.preprocessing import LabelEncoder, StandardScaler\n",
    "from imblearn.over_sampling import SMOTE  # Handling class imbalance\n",
    "\n",
    "# Load the dataset (UCI Credit Risk Dataset)\n",
    "data_url = 'https://archive.ics.uci.edu/ml/machine-learning-databases/00350/default%20of%20credit%20card%20clients.xls'\n",
    "data = pd.read_excel(data_url, header=1)\n",
    "\n",
    "# Drop irrelevant rows if any and reset the index\n",
    "data = data[1:]\n",
    "data.reset_index(drop=True, inplace=True)\n",
    "\n",
    "# Rename columns for consistency\n",
    "data.columns = [\n",
    "    'ID', 'LIMIT_BAL', 'SEX', 'EDUCATION', 'MARRIAGE', 'AGE', 'PAY_0', 'PAY_2', \n",
    "    'PAY_3', 'PAY_4', 'PAY_5', 'PAY_6', 'BILL_AMT1', 'BILL_AMT2', 'BILL_AMT3',\n",
    "    'BILL_AMT4', 'BILL_AMT5', 'BILL_AMT6', 'PAY_AMT1', 'PAY_AMT2', 'PAY_AMT3',\n",
    "    'PAY_AMT4', 'PAY_AMT5', 'PAY_AMT6', 'default'\n",
    "]\n",
    "\n",
    "# Select relevant columns for modeling\n",
    "selected_columns = [\n",
    "    'LIMIT_BAL', 'AGE', 'SEX', 'EDUCATION', 'MARRIAGE', 'PAY_0', 'PAY_2', \n",
    "    'PAY_3', 'PAY_4', 'PAY_5', 'PAY_6', 'BILL_AMT1', 'BILL_AMT2', 'BILL_AMT3',\n",
    "    'BILL_AMT4', 'BILL_AMT5', 'BILL_AMT6', 'default'\n",
    "]\n",
    "data = data[selected_columns]\n",
    "\n",
    "# Fill missing values with mean\n",
    "data = data.fillna(data.mean())\n",
    "\n",
    "# Standardize numeric columns\n",
    "numeric_columns = ['LIMIT_BAL', 'AGE', 'BILL_AMT1', 'BILL_AMT2', 'BILL_AMT3', 'BILL_AMT4', 'BILL_AMT5', 'BILL_AMT6']\n",
    "scaler = StandardScaler()\n",
    "data[numeric_columns] = scaler.fit_transform(data[numeric_columns])\n",
    "\n",
    "# Encode categorical variables\n",
    "categorical_columns = ['SEX', 'EDUCATION', 'MARRIAGE', 'PAY_0', 'PAY_2', 'PAY_3', 'PAY_4', 'PAY_5', 'PAY_6']\n",
    "for col in categorical_columns:\n",
    "    le = LabelEncoder()\n",
    "    data[col] = le.fit_transform(data[col])\n",
    "\n",
    "# Split the dataset into features and target\n",
    "X = data.drop('default', axis=1)\n",
    "y = data['default']\n",
    "\n",
    "# Handle class imbalance using SMOTE\n",
    "smote = SMOTE(random_state=42)\n",
    "X_resampled, y_resampled = smote.fit_resample(X, y)\n",
    "\n",
    "# Split into training and test sets\n",
    "X_train, X_test, y_train, y_test = train_test_split(X_resampled, y_resampled, test_size=0.2, random_state=42, stratify=y_resampled)\n",
    "\n",
    "# Train a Random Forest Classifier with GridSearchCV\n",
    "param_grid = {\n",
    "    'n_estimators': [100, 200],\n",
    "    'max_depth': [None, 10, 20],\n",
    "    'min_samples_split': [2, 5],\n",
    "    'max_features': ['sqrt', 'log2']\n",
    "}\n",
    "grid_search = GridSearchCV(RandomForestClassifier(random_state=42), param_grid, cv=3, scoring='roc_auc', verbose=1, n_jobs=-1)\n",
    "grid_search.fit(X_train, y_train)\n",
    "\n",
    "# Best estimator\n",
    "best_model = grid_search.best_estimator_\n",
    "\n",
    "# Make predictions\n",
    "y_pred = best_model.predict(X_test)\n",
    "y_proba = best_model.predict_proba(X_test)[:, 1]\n",
    "\n",
    "# Evaluate the model\n",
    "accuracy = accuracy_score(y_test, y_pred)\n",
    "roc_auc = roc_auc_score(y_test, y_proba)\n",
    "print(\"Best Parameters:\", grid_search.best_params_)\n",
    "print(\"Accuracy:\", accuracy)\n",
    "print(\"ROC AUC Score:\", roc_auc)\n",
    "print(\"\\nClassification Report:\\n\", classification_report(y_test, y_pred))\n",
    "\n",
    "# Confusion Matrix\n",
    "conf_matrix = confusion_matrix(y_test, y_pred)\n",
    "sns.heatmap(conf_matrix, annot=True, fmt='d', cmap='Blues', xticklabels=['No Default', 'Default'], yticklabels=['No Default', 'Default'])\n",
    "plt.title('Confusion Matrix')\n",
    "plt.xlabel('Predicted')\n",
    "plt.ylabel('Actual')\n",
    "plt.show()\n",
    "\n",
    "# Feature Importance\n",
    "feature_importance = best_model.feature_importances_\n",
    "importance_df = pd.DataFrame({'Feature': X.columns, 'Importance': feature_importance}).sort_values(by='Importance', ascending=False)\n",
    "\n",
    "# Plot Feature Importance\n",
    "plt.figure(figsize=(10, 6))\n",
    "sns.barplot(x='Importance', y='Feature', data=importance_df)\n",
    "plt.title('Feature Importance')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "### **5. Model Performance**\n",
    "\n",
    "The following evaluation metrics are used:\n",
    "\n",
    "#### **Classification Report**:\n",
    "| Metric           | Class 0 (No Default) | Class 1 (Default) |\n",
    "|------------------|----------------------|-------------------|\n",
    "| **Precision**    | 0.85                 | 0.86              |\n",
    "| **Recall**       | 0.86                 | 0.85              |\n",
    "| **F1-Score**     | 0.86                 | 0.85              |\n",
    "| **Support**      | 4673                 | 4673              |\n",
    "\n",
    "#### **Key Metrics**:\n",
    "- **Accuracy**: 85.4%\n",
    "- **ROC AUC Score**: 0.924\n",
    "\n",
    "---\n",
    "\n",
    "### **6. Confusion Matrix**\n",
    "\n",
    "The Confusion Matrix visualizes the model's predictions:\n",
    "\n",
    "\n",
    "| **Actual vs Predicted** | **No Default (0)** | **Default (1)** |\n",
    "|-------------------------|--------------------|-----------------|\n",
    "| **No Default (0)**      | 4024               | 649             |\n",
    "| **Default (1)**         | 711                | 3962            |\n",
    "\n",
    "---\n",
    "\n",
    "### **7. Feature Importance**\n",
    "\n",
    "The top features contributing to the prediction include:\n",
    "1. **LIMIT_BAL** (Credit Limit)\n",
    "2. **PAY_0** (Payment status in the most recent month)\n",
    "3. **AGE**\n",
    "4. **BILL_AMT1** to **BILL_AMT3** (Bill statements)\n",
    "\n",
    "\n",
    "---\n",
    "\n",
    "### **8. Conclusions**\n",
    "\n",
    "- The **Random Forest Classifier** performed well with an **Accuracy of 85.4%** and an **AUC of 0.924**.\n",
    "- **LIMIT_BAL**, **PAY_0**, and **BILL_AMT** are the most important predictors of credit default.\n",
    "- The dataset is balanced, simplifying the model's performance without additional resampling techniques."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base",
   "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": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}