AI-Driven Health Portal: Machine Learning for Chronic Disease Prediction and Personalized Care

Abstract

Chronic disease is on the rise-anaemia, hypertension, diabetes, and hypercholesterolemia- it, thus, demands continuous monitoring and early diagnosis. We present an AI-based Health Portal System integrating machine learning, predictive analytics, telemedicine, and a chatbot assistant for personalized health insights. Trained on more than 10,000 clinical records, the system employs Logistic Regression, Random Forest, and Decision Tree classifiers, with an accuracy of 90.50%-100% and precision over 97%. The anaemia model predicts severity levels and polycythaemia, the hypertension model analyses blood pressure readings, and the cholesterol model detects hyperlipidaemia. The diabetes model had real-world validation and was very high in predictability, achieving 100% accuracy. The RAG-propelled chatbot powered by FAISS indexing enables rapid retrieval of medical data and telemedicine services. The evaluation metrics authenticate the reliability of the models. Future works will include the expansion of the datasets, hyperparameter tuning, merging with deep learning, and conducting clinical validation for AI's great disruptive role in digital health.

Ähnliche Arbeiten