Xing Ke,Chen Liming,Wu Mengjie,et al.Design and implementation of an AI‑empowered informationized storage and retrieval system for traditional Chinese medicine standards[J].Shanghai Journal of Traditional Chinese Medicine,2026,60(8):1-9.
Xing Ke,Chen Liming,Wu Mengjie,et al.Design and implementation of an AI‑empowered informationized storage and retrieval system for traditional Chinese medicine standards[J].Shanghai Journal of Traditional Chinese Medicine,2026,60(8):1-9.DOI: 10.16305/j.1007-1334.2026.z20251126007.
Design and implementation of an AI‑empowered informationized storage and retrieval system for traditional Chinese medicine standards
To develop an artificial intelligence-based processing pipeline for traditional Chinese medicine (TCM) standard information, specifically, an informationized storage and retrieval system for TCM standards. This system aims to achieve structured extraction, database integration, and retrieval of TCM standard documents, thereby advancing the intelligent development of the TCM standardization system.
Methods
2
Based on the structural characteristics of TCM standard documents, information extraction was performed using format conversion tools such as MinerU. The research group innovatively established a processing flow that included seal removal, file type discrimination, and parallel parsing. It adopted a standard backbone extraction method based on finite state machines and a hierarchical strategy to handle highly unstructured domain-specific content. Leveraging large language model (LLM) combined with retrieval-augmented generation (RAG) and domain-specific prompt engineering, the system achieved precise retrieval of standard data, followed by rigorous validation.
Results
2
The developed system for TCM standards supports semantic retrieval via mainstream LLM using RAG. Validation was conducted using over 400 standards, including national, local, and industry standards within the TCM field. The proposed document format conversion and backbone extraction algorithms both achieved accuracy rates exceeding 95%.
Conclusion
2
The AI-driven storage and retrieval system designed in this study effectively meets full-process informatization requirements, ranging from information extraction and database entry to retrieval of TCM standards.
Nguyen T T H , Jatowt A , Coustaty M , et al . Survey of post-OCR processing approaches [J]. ACM Comput Surv , 2021 , 54 ( 6 ): 124 .
Chauhan A . A review on various aspects of MongoDB databases [J]. Int J Eng Res , 2019 , 8 ( 5 ): 580 - 583 .
Shi W , Tan H , Kuang C , et al . Pangu DeepDiver:adaptive search intensity scaling via Open-Web reinforcement learning [PP/OL]. arXiv ( 2025-05-30 )[ 2025-11-01 ]. https://arxiv.org/abs/2505.24332 https://arxiv.org/abs/2505.24332 .
Wang B , Xu C , Zhao X , et al . MinerU: An Open-Source Solution for Precise Document Content Extraction [PP/OL]. arXiv ( 2024-09-27 )[ 2025-11-01 ]. https://arxiv.org/abs/2409.18839 https://arxiv.org/abs/2409.18839 .
Vaswani A , Shazeer N , Parmar N , et al . Attention is all you need [C]//Curran Associates. Advances in Neural Information Processing Systems . NY : Curran Associates , 2017 : 1 - 11 .
Salemi A , Zamani H . Evaluating retrieval quality in retrieval-augmented generation [C]//Association for Computing Machinery. Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval . NY : Association for Computing Machinery , 2024 : 2395 - 2400 .
Lewis P , Perez E , Piktus A , et al . Retrieval-augmented generation for knowledge-intensive NLP tasks [C]//Neural Information Processing Systems Foundation. Advances in Neural Information Processing Systems . NY : Curran Associates , 2020 : 9459 - 9474 .
Aumüller M , Bernhardsson E , Faithfull A . ANN-Benchmarks: A benchmarking tool for approximate nearest neighbor algorithms [J]. Inf Syst , 2020 , 87 : 101374 .
DeepSeek-AI , Liu A , Feng B , et al . DeepSeek-V3 technical report [PP/OL]. arXiv ( 2024-12-27 )[ 2025-11-01 ]. https://arxiv.org/abs/2412. 19437 https://arxiv.org/abs/2412.19437 .
Yang A , Li A , Yang B , et al . Qwen3 technical report [PP/OL]. arXiv ( 2025-05-14 )[ 2025-11-01 ]. https://arxiv.org/abs/2505.09388 https://arxiv.org/abs/2505.09388 .
Systematic construction strategy for preclinical efficacy evaluation platform of Chinese materia medica compounds
Quantitative study on core pathogenesis of chronic heart failure in traditional Chinese medicine based on fuzzy cognitive maps
Development of diagnostic model for liver depression and spleen deficiency syndrome in colorectal cancer patients during chemotherapy based on machine learning
Construction of a predictive model for type 2 diabetes remission based on artificial intelligence integration of traditional Chinese medicine tongue manifestations, syndrome elements and Western medical indicators
Application of machine learning algorithms in screening of active components in Chinese materia medica
Related Author
Yao Yuqin
Wang Yu
Yang Yang
Ji Xinke
Zhao Yun
Zhang Xueqin
Li Ping
Wen Tiancai
Related Institution
Science and Technology Experimental Center, Shanghai University of Traditional Chinese Medicine
National Data Center of Traditional Chinese Medicine, China Academy of Chinese Medical Sciences
Department of Cardiology, Beijing University of Chinese Medicine Third Affiliated Hospital
Department of Oncology, Cangzhou Hospital of Integrated Traditional Chinese and Western Medicine
Department of Oncology, Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine