Article Details

Abstract

Aiming at the problems of low efficiency and insufficient accuracy in retrieval and generation caused by the multi-source heterogeneity and semantic complexity of domestic medical data, this paper proposes a Retrieval-Augmented Generation method for medical event tags (ME-RGA). This method integrates large language models and RAG technology. Through visual tag design, it extracts data with a “structured + large model” dual-drive approach, and constructs a tag library through text chunking and vectorization processing. Relying on AI analysis, it completes input vectorization, similarity matching, semantic retrieval, and large model reasoning. Experiments show that this method significantly improves retrieval efficiency and generation accuracy, can provide high-quality support for clinical decision-making, and has important application value and broad development prospects.

Keywords

Large language model; Retrieval-Augmented Generation; Prompt Engineering; medical event tag

References

1. [1] Zhong Guixin, Li Jiagen, He Xiong, etc Exploration of the Application of Large Language Models in Clinical Medicine at Home and Abroad [J]. Medical Theory and Practice, 2025, 38 (14): 2371-23742385
2. [2] Li Xiaotian Design and Practice of Data Labeling System [J]. Electronic Technology, 2024, 53 (8): 282-283
3. [3] Wang Xia, Xu Xiangdong, Zhou Guanghua, etc Research on the Construction Method of Health and Medical Big Data Label System [J]. Chinese Journal of Health Information Management, 2021, 18 (2): 189-193
4. [4] Dai Min, Zhu Sen, Lv Xiaoyun A method for calculating semantic similarity of medical terms based on SNOMED-CT [J]. Journal of Tianjin University of Technology, 2021, 37 (6): 27-30, 52
5. [5] Valenzuela-Escarcega M A，Hahn-Powell G，Surdeanu M，et al. A Domain-independent Rule-based Framework for Event Extraction [C]//Proceedings of ACL-IJCNLP 2015 System Demonstrations，2015：127-132.
6. [6] Lu Yuting Medical guideline events and event relationship extraction based on syntactic analysis [D]. Wuhan: Wuhan University of Science and Technology, 2014
7. [7] Tian Z，Sun S，Eguale T，et al. Automated Extraction of VTE Events From Narrative Radiology Reports in Electronic Health Records：A Validation Study [J].Medical Care，2017，55（10）：e73-e80.
8. [8] Tommaso C，Roser M. VUACLTL at SemEval 2016 Task 12: A CRF Pipeline to Clinical TempEval [C]//Proceedings of the 10th International Workshop on Semantic Evaluation (SemEval-2016)，2016：1241-1247.
9. [9] Zeng Lu, Gao Daqi, Ruan Tong, etc Analysis and annotation of symptom composition based on CRF [J]. Journal of East China University of Science and Technology (Natural Science Edition), 2018, 44 (2): 277-282
10. [10] Negi K，Pavuri A，Patel L，et al. A Novel Method for Drug-adverse Event Extraction Using Machin Learning [J/OL].Informatics in Medicine Unlocked，2019，17：100190.https://doi.org/10.1016/j.imu.2019.100190.
11. [11] Ramponi A，Plank B，Lombardo R. Cross-domain Evaluation of Edge Detection for Biomedical Event Extraction [C]//Proceedings of the 12th Language Resources and Evaluation Conference，2020：1982-1989.
12. [12] Xie J D，He J Y，He W M，et al. Research on Structured Information Extraction Method of Electronic Medical Records of Traditional Chinese Medicine [C]//2020 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)，2020：1613-1616.
13. [13] Xing D W，Ulf L，Beeds L W. BEEDS: Large-Scale Biomedical Event Extraction using Distant Supervision and Question Answering [C]//Proceedings of the 21st Workshop on Biomedical Language Processing，2022：298-309.
14. [14] Li L S，Liu Y，Qin M Y. Extracting Biomedical Events with Parallel Multi-Pooling Convolutional Neural Networks [J].IEEE-ACM Transactions on Computational Biology and Bioinformatics，2020，17（2）：599-607.
15. [15] Guo Yujie The Design and Implementation of the System for Building the Medical Atlas Based on Electronic Medical Record [D]. Shenyang: Chinese Academy of Sciences University (Chinese Academy of Sciences Shenyang Institute of Computing Technology), 2022
16. [16] Yue Shuhong Research on Event Extraction for Chinese Electronic Medical Records [D]. Taiyuan: Shanxi University, 2023
17. [17] Xing D W，Leon W，Ulf L. Biomedical Event Extraction as Multi-turn Question Answering [C]//Proceedingsof the 11th International Workshop on Health Text Mining and Information Analysis，2020：88-96.
18. [18] Ni Hao, Zhang Xu, Ma Shuai, etc Electronic Medical Record Medical Event Extraction Model Based on CasEE [J]. Journal of Xiamen University of Technology, 2023, 31 (3): 40-49
19. [19] Yan S K，Wong K C. Context awareness and embedding for biomedical event extraction [J].Bioinformatics，2020，36（2）：637-643.
20. [20] Yu Jie, Ji Bin, Liu Lei, etc Joint Extraction Method for Chinese Medical Events [J]. Computer Science, 2021, 48 (11): 287-293
21. [21] Gao Su, Tao Hu, Jiang Yanzhao, etc Sentence level Joint Event Extraction of Traditional Chinese Medicine Literature [J]. Information Engineering, 2021, 7 (5): 15-29
22. [22] Cao Yaya Research on Event Joint Extraction Method Based on LERT Pre trained Model [D]. Changchun: Changchun University of Technology, 2024
23. [23] Wu T T，Shiri F，Kang J Q，et al. KC-GEE: Knowledge-Based Conditioning for Generative Event Extraction [J].World Wide Web: Internet and Web Information Systems，2023，26（6）：3983-3999.
24. [24] Wang Lu, Ding Mufei, Zhou He, etc A review of the development and application system of medical language models [J]. Journal of Intelligent Systems, 2025, 20 (6): 1295-1303
25. [25] Touvron H，Lavril T，Izacard G，et al. LLaMA: Open and Efficient Foundation Models [PP/OL].https://arxiv.org/abs/2302.13971.
26. [26] Wu Xuan, Fu Tao A review of research on retrieval enhanced generation technology [J]. Computer Engineering and Applications, 2025, 61 (20): 19-35
27. [27] Huang Jun, Lin Fei, Yang Jing, etc The Large Model Hint Engineering of Generative AI: Methods, Current Status, and Prospects [J]. Journal of Intelligent Science and Technology, 2024, 6 (2): 115-133
28. [28] Wang Dongqing, Lu Fei, Zhang Binghui, etc Overview of Hint Word Engineering in Large Language Models [J]. Computer System Applications, 2025, 34 (1): 1-10
29. [29] Du Jiale, Chen Shudong, Ye Liang, etc Overview of Thinking Chain Techniques in Large Language Models [J]. Radio Communication Technology, 2025, 51 (5): 877-887

Funding

National Science and Technology Major Special Project "Selection, Integration, and Evaluation of Clinical Diagnosis, Treatment, Emergency and Rehabilitation Technologies" (2024ZD0524202)

About the authors

Zhang Sumin (1981-), female, Han ethnicity, born in Taizhou, Jiangsu, is an engineer and a master's student. Her research interests include health information management and large-scale modeling; Corresponding author: Shi Yan (1985-), male, Han nationality, from Zhumadian, Henan, senior engineer, bachelor's degree, research direction: health information management, artificial intelligence, blockchain.