A machine learning–driven project to predict loan default risk and generate transparent credit scores.
The Automated Credit Score Evaluation System employs a logistic regression model to:
- Estimate the probability that a loan applicant will default.
- Map that probability to a credit score between 300 and 900.
- Categorize risk levels: Poor (300–500), Fair (501–650), Good (651–800), Excellent (801–900).
This tool provides both batch processing for large datasets and an interactive UI for single-customer scoring.
- High-Recall Default Prediction: Optimized to minimize false negatives on defaulters.
- Credit Score Mapping: Customizable scoring curve following industry norms.
- Risk Tiers: Intuitive categories for decision-making.
- Data Leakage Prevention: Strict train/test split before feature transformations.
- Class Imbalance Handling: SMOTE Tomek pipeline ensures balanced training.
- Feature Engineering: Ratios like Loan-to-Income (LTI), delinquency frequency, and average days past due.
- Hyperparameter Optimization: Optuna-driven tuning for best performance.
- Evaluation Metrics: AUC-ROC, Gini (> 0.7), KS statistic (> 0.4), decile-wise analysis.
- Interactive Streamlit App: Instant single-user scoring with probability plots.
https://loan-defaulter-score-system.streamlit.app/
app/
│
├── artifacts/
│ ├── model_data.joblib # Trained ML model
│
├── main.py # Core Streamlit app
├── prediction_helper.py # Scoring logic & preprocessing helpers
├── app_image.jpg # Local image for README
├── .gitignore # Git ignore rules
├── requirements.txt # Project dependencies
└── README.md # Project overview
-
Clone the Repository
git clone https://github.com/MindMatrixPro/Automated_Credit_Score_System.git cd credit-score-system -
Install Dependencies
pip install -r requirements.txt
-
Run the Streamlit App
streamlit run main.py
To Contribute, Please submit issues or pull requests for enhancements or fixes.
Licensed under the Apache 2.0 License.
Transform credit risk management with precision and ease!