Credit Card

💳 Credit Card Approval Prediction

📂 GitHub Repository

🔗 Link to GitHub Repo

📌 Project Overview

This project leverages machine learning to predict whether a credit card application will be approved or rejected based on user demographics and financial details. By analyzing key customer attributes such as income, employment status, and credit history, the model helps automate the approval process with high accuracy.

📂 Dataset

• Source: Kaggle
• Target Column: Status (Approved/Rejected)

🔍 Steps Involved

1. Data Cleaning – Handled missing values and inconsistencies.
2. Exploratory Data Analysis (EDA) – Discovered key patterns affecting approvals.
3. Feature Selection – Identified the most important predictors.
4. Data Encoding – Converted categorical variables into numerical formats.
5. Model Training & Evaluation – Tested multiple ML models to find the best performer.

🚀 Model Performance

Model	Accuracy	F1 Score	Precision	Recall
Random Forest	98.76%	98.48%	98.30%	98.76%
XGBoost	98.76%	98.34%	97.98%	98.76%
SVM	98.91%	98.37%	97.84%	98.91%

🔹 Why Random Forest?

Although SVM achieved the highest accuracy (98.91%), Random Forest was chosen as the final model due to its:

• High F1-Score (98.48%), ensuring a balance between Precision & Recall.
• Robustness against overfitting compared to complex models.
• Ability to handle both categorical & numerical data efficiently.

📌 Future Improvements

• Deploy the model as a web app for real-time predictions.
• Enhance interpretability using SHAP (SHapley Additive Explanations).
• Integrate real-world financial data for better generalization.