Keywords
Knowledge graph
AGBAN model
Inheritance of famous TCM physicians’ experience
Clinical reasoning
Submitted : 2026-04-22
Accepted : 2026-05-07
Published : 2026-05-22
Abstract
Aiming at the problems of knowledge fragmentation and opaque reasoning in the digital inheritance of famous TCM physicians’ academic thoughts and diagnosis-treatment experience, a multimodal knowledge graph construction method based on the AGBAN model is proposed. Using more than 3,000 outpatient medical records of famous TCM physicians as the data source, multimodal information is integrated to construct a clinical knowledge graph through ontology design, entity-relationship extraction, and knowledge storage. The graph attention network and reinforcement learning mechanism of the AGBAN model are introduced to optimize the diagnosis-treatment path. The results show that the knowledge graph contains 3,089 entities and 1,461 relationships, with an average degree of 2.49; the average reciprocal rank of link prediction of the AGBAN model is 0.973, which is 165.4% higher than that of the TransE model, the diagnosis success rate is 59.19%, and the average reasoning path is 5 steps; cluster analysis verifies the core TCM principles such as “drug-syndrome correspondence”. The conclusion indicates that this method realizes the structured representation and intelligent reasoning of famous TCM physicians’ clinical experience, providing a feasible path for TCM academic inheritance and clinical decision support.
References
Xiang X, 2018, Preliminary Exploration of the Experience in the Inheritance of TCM Academic Schools and Famous and Senior TCM Physicians. In: The 5th Lanmao Forum and 2018 Yunnan Provincial TCM Academic Annual Conference. Yunnan Provincial Association of Traditional Chinese Medicine et al.
China Traditional Chinese Medicine News, 2022, Qihuang Project Cultivates TCM Talents—Selecting 3,404 Instructors and Training 6,562 Successors, the National Inheritance Work of Academic Experience of Famous and Senior TCM Experts Achieves Great Results.
Yang H, Wang J, 2019, Construction and Visualization of the Knowledge Graph of Academic Inheritance in the Song Dynasty. New Technology of Library and Information Service, 3(6): 109–116.
Chen Y, Xie W, Chen F, et al., 2024, Discussion on the Construction and Visualization of the Knowledge Graph of TCM Ancient Book Lingshu. Journal of Guangzhou University of Chinese Medicine, 2024(3): 41.
Zheng C, Wu Z, Wang T, et al., 2021, Object-Aware Multimodal Named Entity Recognition in Social Media Posts with Adversarial Learning. IEEE Transactions on Multimedia, 23(10): 2520–2532.
Pandian J, Thirunavukarasu R, Nagarajan R, 2025, Enhanced Exploration in Reinforcement Learning Using Graph Neural Network Based Intrinsic Reward Mechanism. Scientific Reports, 15(1): 39986.
Summer G, Kelder T, Ono K, et al., 2015, CyNeo4j: Connecting Neo4j and Cytoscape. Bioinformatics, 31(23): 3868–3869.
Konstantinov A, Kozlov B, Kirpichenko S, et al., 2025, Dual Feature-Based and Example-Based Explanation Methods. Frontiers in Artificial Intelligence, 8: 1506074.
Diaz-Roldan C, Prats M, Ramos-Herrera M, 2021, Redefining Monetary Policy Rules: A Threshold Approach. PLoS ONE, 16(5): e0252316.
Francis N, Green A, Guagliardo P, et al., 2018, Cypher: An Evolving Query Language for Property Graphs. In: Proceedings of the 2018 International Conference on Management of Data (SIGMOD). 1433–1445.
Gu S, Zhang G, Jia C, et al., 2025, Attention-Based Batch Normalization for Binary Neural Networks. Entropy, 27(6): 645.
Bennett C, Hauser K, 2013, Artificial Intelligence Framework for Simulating Clinical Decision-Making: A Markov Decision Process Approach. Artificial Intelligence in Medicine, 57(1): 9–19.
Bordes A, Usunier N, Garcia-Duran A, et al., 2013, Translating Embeddings for Modeling Multi-Relational Data. In: Advances in Neural Information Processing Systems. 1–9.
Ma X, Wang M, Ma J, et al., 2022, The Association Between Lifestyles and Health Conditions and the Choice of Traditional Chinese Medical Treatment in China: A Latent Class Analysis. Medicine, 101(51): e32422.
Bai Y, Du J, Li H, et al., 2025, An Adaptive Framework for Remaining Useful Life Prediction Integrating Attention Mechanism and Deep Reinforcement Learning. Sensors, 25(20): 6354.