Ayurvedic Medicine Recommendation

Ayurveda, an ancient Indian medical system, offers personalized healthcare through the assessment of Prakriti (constitution), dosha imbalance, symptoms, and lifestyle factors. However, Ayurvedic diagnosis and medicine selection often depend heavily on expert practitioner interpretation, making it difficult to scale and standardize. With advances in artificial intelligence (AI) and machine learning (ML), personalized recommendation systems can improve the accessibility, consistency, and efficiency of Ayurvedic healthcare. This paper presents a machine learning–based framework for Ayurvedic medicine recommendation using structured patient data, dosha assessment, and symptom analysis. A standardized ontology for Ayurvedic concepts (diseases, herbs, formulations, and dosha associations) was developed to address heterogeneity in classical terminology. The proposed system uses supervised learning models—Decision Trees, Random Forests, and Support Vector Machines (SVM)—to predict appropriate Ayurvedic formulations. An NLP-based component extracts therapeutic associations from Ayurvedic texts. Experiments on a curated dataset of 1,500 patient records show that Random Forest achieves the highest accuracy (94.8%) for medicine recommendation. The study highlights the potential of ML to augment Ayurvedic clinical decision support, while also discussing limitations and ethical issues.