Credit risk ML pipeline with CatBoost, SHAP & LIME explainability, fairness monitoring and auto generated PDF reports built for auditability over accuracy.
-
Updated
Mar 21, 2026 - Python
Credit risk ML pipeline with CatBoost, SHAP & LIME explainability, fairness monitoring and auto generated PDF reports built for auditability over accuracy.
A practical exploration of Interpretable Machine Learning (XAI) with Python. This repository features Jupyter notebooks covering SHAP, LIME, Anchors, Counterfactuals, Grad-CAM, and more, applied to models from Linear Regression to Convolutional Neural Networks (CNNs).
企业员工离职风险预警系统(HRA) - LightGBM + IsolationForest + SHAP 多模态融合 · 测试AUC 0.9862 · FastAPI + Vue3 · 漂移/公平性治理 + 2FA/PII加密 · 288+ tests
Rank heart-disease models by clinical cost, not accuracy: the 5th-most-accurate saves $9.7M per 10K patients: cost-sensitive model selection, fairness & robustness audits
FastAPI deployment exercises plus a model fairness/bias audit (Aequitas) on a COMPAS-style dataset.
To associate your repository with the model-fairness topic, visit your repo's landing page and select "manage topics."