Multimodal AI Models for Early Detection of Chronic Diseases Using Lifestyle and Biometric Data

Early detection of chronic diseases remains a major challenge in modern healthcare. We present a multimodal AI framework that fuses biometric, lifestyle, and environmental data from wearables, clinical records, and sensor networks to predict the onset of conditions such as Type ۲ diabetes, hypertension, and cardiovascular disease. The system combines XGBoost for structured features, LSTM for temporal lifestyle signals, and Graph Neural Networks to model relational health dependencies, prioritizing early-stage sensitivity, interpretability, and personalization. Evaluation on real-world datasets shows improved accuracy and sensitivity compared with single-source baselines, with lifestyle integration substantially enhancing early detection. Future work will focus on real-time deployment via wearable devices and integration with national health systems to enable equitable, targeted preventive care.

کلیدواژه ها:

Artificial Intelligence ، Chronic Disease Prediction ، Multimodal Data ، Machine Learning ، Healthcare AI