The goal was to build a predictive model that accurately identifies which liability customers are most likely to accept a personal loan offer. I approached this through supervised classification using two key models: Logistic Regression and Decision Tree.

🧹 Data Preparation

• Target Variable: Personal_Loan (0 = No, 1 = Yes)
• Dropped: ID, ZIPCode (not predictive)
• Feature Scaling: Standardized continuous variables (e.g., Income, CCAvg, Mortgage) using StandardScaler
• Train-Test Split: 70% train, 30% test (stratified on target)

📦 Model 1: Logistic Regression

Why this model?

• Great for binary classification problems
• Highly interpretable coefficients
• Strong performance when data is linearly separable

Results:

• Accuracy: ~96%
• ROC AUC Score: ~0.95
• Precision/Recall: High precision and recall for positive class

🔍 Interpretation: The odds of accepting a loan increase significantly with higher income, credit card usage, and CD account ownership.

🌳 Model 2: Decision Tree Classifier