Disease Prediction System using Machine Learning

The Chronic Disease Prediction System is an AI-powered web platform developed to enable early and accurate screening for pneumonia, stroke, and diabetes in a single unified environment. Leveraging deep learning and machine learning techniques, it integrates a convolutional neural network for chest X-ray analysis, logistic regression for stroke assessment, and a random forest classifier for diabetes detection, thereby accommodating both image and tabular clinical data. The project features robust user authentication and personalized dashboards, allowing secure access and tracking of results. For pneumonia detection, the system not only yields a probability score but also visualizes affected lung regions using Grad- CAM heatmaps, supporting rapid triage and clinical interpretation. Stroke and diabetes modules accept structured health inputs, offering risk classifications with pie chart visualizations and direct links to medical consultations. Its scalable architecture, built on Flask (backend) and React (frontend), is designed for cross-platform accessibility and future extensibility, enabling integration with additional diseases, cloud deployment, and advanced explainability features like SHAP and LIME. By combining rapid inference, consistency, user-centric design, and health professional integration, the system advances preventive medicine, offering considerable impact in cost reduction, accessibility, and clinical workflow efficiency.