CAD Detection using WGAN-GP and Transformer with Explainability

This project presents a deep learning pipeline for Coronary Artery Disease (CAD) detection using cytokine biomarkers. The methodology includes:- Data preprocessing using KNN imputation and normalization- WGAN-GP for synthetic minority data generation- Transformer-based classification model- Explainability techniques including attention analysis, permutation importance, and integrated gradients The package includes:- Synthetic cytokine dataset- Complete Python implementation- Reproducibility files (requirements, README) This work supports reproducible AI research in medical data analysis.