AI-Driven Ayurvedic Chatbot: Revolutionizing Formulation Selection and Knowledge Accessibility in Traditional Medicine Through the Integration of LENET and Stacked Autoencoders

Ayurveda's vast medicinal knowledge, scattered across numerous texts, presents a formidable challenge for practitioners and students. To address this, we propose an innovative Ayurvedic Chatbot leveraging Artificial Intelligence (AI) and Natural Language Processing (NLP) technology. The proposed model is trained with a fusion of LeNet for feature extraction and Stacked Autoencoders (SAEs) for textual and pharmacological data analysis, enabling the chatbot to streamline the identification of tailored formulations. LeNet processes data aspects, recognizing patterns in input query, while stacked autoencoders classifies the disease symptoms and suggests the ayurvedic medical combination details. Beyond formulation recommendations, the chatbot navigates patient-specific nuances, considering contraindications and suitable mediums for specific diseases. It tackles Ayurvedic terminology intricacies, clarifying synonymous terms and disease name variations. After rigorous experimentation and validation using a medical dataset, the proposed model demonstrated remarkable performance with an accuracy score of 0.97, precision reaching 0.95, recall standing at 0.94, and an F1 score of 0.94. The proposed research not only revolutionizes formulation selection but also serves as an educational tool, providing comprehensive insights for practitioners and students. With a user-friendly interface, this Ayurvedic Chatbot redefines accessibility to Ayurvedic knowledge, offering precise, personalized recommendations for holistic healthcare interventions within the rich Ayurvedic tradition.