Traditional Chinese Medicine (TCM) relies fundamentally on personalized “syndrome differentiation,” yet transitioning this clinical precision to post-treatment medicine selection remains a significant challenge in digital environments. In typical e-commerce settings, recommendation engines often lack personalization, relying instead on generic best-seller lists or simplistic symptom-matching that fails to leverage the wealth of patient-specific data available in TCM clinics. This recognized gap often leads to patient non-adherence and suboptimal health outcomes, as existing systems face technical hurdles such as the “cold-start” problem, where collaborative filtering fails new patients, and a lack of clinical intelligence in pure content-based filtering (Ye et al., 2022). This study addresses these issues by proposing a hybrid algorithm that integrates Natural Language Processing (NLP) for symptom analysis with machine learning techniques like k-nearest neighbours (K-NN) to identify similar patient profiles. By dynamically weighing clinical health records against digital purchase behaviours, the system ensures transparency through Explainable AI (XAI) and maintains ethical integrity through data anonymization. Ultimately, this research introduces a novel framework that empowers TCM clinics to provide clinically aligned, trustworthy product suggestions, bridging the gap between traditional healing wisdom and modern data-driven e-commerce to improve patient adherence and retention.

Keywords: personalized treatment, Traditional Chinese Medicine (TCM), hybrid recommendation algorithm,patient health records, collaborative filtering