Construction and Applications of a Traditional Chinese Medicine Knowledge Graph

The report introduces the need for organizing the massive, fragmented knowledge in the field of Traditional Chinese Medicine (TCM) and proposes knowledge‑graph technology as a solution for systematic structuring, integration, and intelligent services such as recommendation and question answering.

It describes the definition of a TCM knowledge graph, emphasizing its semantic‑network core, inclusion of synonyms, definitions, attributes, and links, and explains why such a graph is essential for AI‑driven TCM applications.

The construction workflow is outlined in four main steps: top‑level design with semantic standards, building the semantic network skeleton, importing relational and semi‑structured data, and expanding the graph through crowdsourced processing and text‑knowledge extraction.

Key components of the graph’s foundation are detailed, including the TCMLS semantic standard (type system, relationships, and ontology), the hierarchical top‑level framework, and the extensive lexical base containing roughly 120,000 concepts and 300,000 terms.

Various sub‑graphs are presented—TCM health, clinical, herb, famous‑doctor inheritance, and specialty therapy—showing how they interlink to form a comprehensive system.

Application scenarios are explored: enhanced knowledge retrieval using synonym and related‑entity expansion, visualisation of concept relationships, intelligent recommendation of herbs, formulas, and therapies, and question‑answering systems that translate natural‑language queries into graph queries.

Specific case studies illustrate the clinical knowledge graph for spleen‑stomach diseases, the processing pipeline from medical case texts to graph entities, and the famous‑doctor inheritance graphs that capture lineage, expertise, and treatment patterns.

The TCM health knowledge graph is used for personalised wellness recommendations based on constitution types, linking lifestyle, diet, and therapeutic methods to individual traits.

In conclusion, the TCM knowledge graph achieves visualisation, integration, and intelligent utilisation of TCM knowledge, supporting retrieval, recommendation, QA, and clinical decision‑support systems, and demonstrates the broader value of graph‑based AI in the medical domain.