Credit Card Approval Prediction

🏦 Credit Card Approval Prediction

A machine learning project that predicts whether a credit card applicant is a good or bad debtor based on financial and demographic data.

📁 Dataset

Two datasets are used:

application_record.csv — applicant demographic and financial info

credit_record.csv — monthly credit payment history per applicant

🔄 Project Pipeline

1. Data Preprocessing

Merged application and credit history datasets on applicant ID

Engineered Age and WorkingYears from raw date fields

Dropped missing values and removed duplicate records

Encoded categorical features using Label Encoding

2. Target Label Creation

Classified applicants as good or bad based on their debt payment history

Applicants with no credit record were labeled using an income threshold

3. Handling Class Imbalance

Applied SMOTE to oversample the minority class and balance the dataset

4. Models Trained

Model Tuned

Logistic Regression ✗

Random Forest ✅ Optuna

Gradient Boosting ✗

XGBoost ✗

MLP Neural Network ✗

5. Evaluation Metrics

Accuracy, Precision, Recall, F1-Score, ROC-AUC

Confusion Matrix & ROC Curve comparison across all models

🏆 Results

Random Forest achieved the best performance after Optuna tuning

Ensemble methods consistently outperformed linear models

Top features: Total Income, Age, Working Years, Family Size