Web-Integrated Machine Learning System For Multi-Disease Risk Assessment

Abstract

Current health screening systems generally focus on identifying single diseases, requiring patients to undergo multiple independent tests and evaluations. This fragmented approach reduces efficiency and accessibility. To address this limitation, we developed a web-integrated machine learning platform capable of performing simultaneous multi-disease risk assessment.

The proposed system evaluates the risk of four major diseases: Diabetes, Thyroid Disorder, Heart Disease, and Kidney Disease using ensemble machine learning models, specifically XGBoost and Random Forest. The models were trained on clinical datasets with proper data preprocessing techniques including feature normalization and cross-validation to ensure high accuracy and reliability.

The platform integrates a FastAPI-based backend for model deployment and a React-based web interface that provides real-time predictions through an interactive dashboard. This solution enhances accessibility for both healthcare professionals and the general public by enabling quick, efficient, and consolidated disease risk screening within a single platform.