AI-DRIVEN PRAKRITI ASSESSMENT AND AYURVEDIC DRUG RECOMMENDATION SYSTEM

Ayurveda emphasizes personalized treatment based on an individual's Prakriti, but conventional assessment relies on manual questionnaires and expert judgment, which are subjective, time-consuming, and difficult to scale.This paper presents HerbiSense, a hybrid machine learning-based system for automated Prakriti assessment and personalized Ayurvedic drug recommendation.The framework integrates Random Forest for Prakriti classification, K-Means clustering for patient profiling, Natural Language Processing (NLP) for symptom interpretation, XGBoost for drug effectiveness prediction, and an Ayurvedic rule-based engine to ensure guideline-compliant decisions.User inputs related to physical, physiological, and behavioral traits, along with free-text symptoms, are collected via a web-based chatbot, preprocessed, and analyzed through this hybrid pipeline.Experimental evaluation shows that the system achieves high Prakriti classification accuracy, generates consistent and relevant recommendations, and responds within acceptable time limits, thereby reducing subjectivity and improving scalability compared to traditional consultation.These results indicate that combining classical Ayurvedic knowledge with modern machine learning can support data-driven, interpretable, and personalized Ayurvedic healthcare