Explainable Healthcare Diagnostics App

A Dual-Mode ML Platform for Diabetes Risk Prediction
This project develops a machine learning-based diabetes risk prediction system using Explainable AI (SHAP/LIME), served through one application built for two distinct users: clinicians and patients. Rather than a one-size-fits-all interface, the app offers a Dual-Mode: Clinician Mode shows dense, metric-driven data (SHAP/LIME plots) for fast diagnosis, while Patient Mode simplifies the same output into plain language (e.g., "Morning Blood Sugar" instead of "Fasting Plasma Glucose") with a simple risk-level indicator. Objectives: build an accurate classifier, integrate dual-audience explainability, and design mode-specific UX. Methodology covers data preprocessing, model training, XAI integration, and a shared backend serving two front-ends. Expected outcomes: a validated predictive model, a jargon-free patient explanation layer, and a professional clinician view bridging clinical utility and everyday accessibility in one platform.