Interpretable and Explainable Classification for Medical Data
📂 Source code unavailable due to collaboration restrictions.
This project explores interpretable and explainable machine learning techniques applied to both tabular and image-based data.
We implemented and evaluated several models including Logistic Regression with Lasso regularization, Multi-Layer Perceptrons (MLPs), Neural Additive Models (NAMs) and Convolutional Neural Networks (CNNs). Also investigated and compared a range of explainability methods, including SHAP, Integrated Gradients, and Grad-CAM, aiming to provide insight into model decisions. Results were compared across models and explanation techniques.
Below, you’ll find some results visualizations from our reports.
