Abstract

Traditional Chinese Medicine (TCM) possesses a profound theoretical framework and extensive clinical history, yet its modernization is frequently impeded by unstructured data, terminological ambiguity, and heterogeneous information sources. Knowledge Graph (KG) technology, serving as an advanced structured representation tool, offers a robust solution for the digitalization and systematic reorganization of TCM knowledge. This review systematically investigates the construction, application, and integration of TCM KGs with Large Language Models (LLMs). Initially, it analyzes the core characteristics of TCM data and the practical challenges of informatization, establishing the necessity for domain-specific KG development. Subsequently, the paper details the foundational architecture of TCM KGs, encompassing critical processes such as ontology design, knowledge extraction, data fusion, storage, and logical reasoning. It further evaluates the current research landscape regarding TCM KGs in pharmacological studies, disease diagnosis, the preservation of ancient classics, and diverse clinical scenarios. The application value across clinical treatment, health management, and educational research is comprehensively summarized. Finally, the review explores the synergistic relationship between LLMs and TCM KGs, assessing practical integration achievements while identifying technical adaptability issues and engineering costs associated with deploying general LLMs in specialized medical fields. Ultimately, the in-depth integration of TCM KGs with domain-specific LLMs represents a pivotal future direction, promising to advance the inheritance, global communication, and modernization of traditional medical knowledge.

Keywords: 

traditional chinese medicine, knowledge graphs, large language models, knowledge construction, knowledge services

License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

How to Cite