Skip to content
0.94 ROC-AUC

Credit Risk + SHAP

An XGBoost risk model that explains every prediction.

Technology Stack

  • Python
  • XGBoost
  • Scikit-learn
  • SHAP

Highlights

  • XGBoost classification pipeline with preprocessing and 5-fold cross-validation: 92% accuracy and 0.94 ROC-AUC.
  • SHAP for global and per-prediction explainability.
  • Hyperparameter tuning, calibration, and error analysis to interpret risk predictions.

“Excellence is a continuous process and not an accident.”

~ A. P. J. Abdul Kalam