An Interpretable Decision Tree Model for Ayurvedic Drug Recommendation in Personalized Healthcare

This study explores the integration of traditional Ayurveda with contemporary machine learning techniques to develop a predictive model for personalized drug recommendation. Using a dataset comprising 1,300 patients samples that detail diseases, demographics, severity levels and corresponding Ayurvedic treatments, a Decision Tree Classifier-based model was constructed to predict Ayurvedic drug prescriptions. Key preprocessing steps, including handling duplicates, null values, and categorical encoding, ensured the data's integrity and usability. The model's predictive capabilities were evaluated against alternative algorithms such as Random Forest, SVM, Logistic Regression, and K-Nearest Neighbors, demonstrating superior accuracy of 95.3% with Decision trees on the test dataset. The proposed system provides an interpretable and robust framework for recommending Ayurvedic treatments based on patient-specific features, emphasizing its clinical applicability and trustworthiness.