A Comprehensive System for Diabetes Management and Prediction Using AI-Driven Chatbot and Machine Learning Models

The integration of a chatbot with a predictive analytics model for Better Diabetes Management, powered by AI. The chatbot has been developed to answer specific user questions using Gradio and the Llama Index with the HuggingFace models in relation to the prevention, care, and management of diabetes in a personalized and evidence-based manner. It does so while making references dynamically to medical documents ingested and thereby assuring relevance in context and up-to-date grounding in medical knowledge. In parallel, the predictive model uses the Flask web application deployment of a Random Forest algorithm to evaluate the risk for diabetes based on key health metrics: glucose levels, blood pressure, and BMI. Inputs are then preprocessed and scaled before analysis, which lets the model provide accurate predictions about the presence of diabetes and gives the user immediate actionable feedback. This integrated system does not only encourage interactive engagement to better understand diabetes, but also provides real-time insights for informed decisions. It is a bridge project between expert medical knowledge and available health management tools in the domain of personalized health through AI.