Library and Information Service

Library and Information Service >

2026 , Vol. 70 >Issue 6: 30 - 39

DOI: https://doi.org/10.13266/j.issn.0252-3116.2026.06.003

SPECIAL TOPIC: AI Enables the Inheritance and Innovation of Herbal Classics

Research on Joint Extraction of Entity Relations of Herbal Classics Based on BSTR-GPRel

  • Shen Wang ,
  • Sun Ke ,
  • Ma Ruolin ,
  • Li Xin ,
  • Bian Wei
Expand
  • School of Business and Management, Jilin University, Changchun 130012
Shen Wang, Professor, PhD, Doctoral Supervisor, E-mail: shenwang_001@163.com; Sun Ke, Master’s Candidate; Ma Ruolin, Master’s Candidate; Li Xin, Master’s Candidate; Bian Wei, Associate Research Fellow, Master’s Degree.

Received date: 2025-06-25

  Revised date: 2025-09-15

  Online published: 2026-03-14

Supported by

This work is supported by the Special Project of Unpopular Extinctive Subjects of the National Social Science Found of China, titled “Research on the Collection, Knowledge Organization and Intelligentization Construction of Classical Herb Books” (Grant No. 23VJXT024).

Fold

Abstract

[Purpose/Significance] This study aims to promote the knowledge structurization and intelligent application of herbal classics, and to support the digital preservation of the cultural heritage of traditional Chinese medicine. [Method/Process] This study proposes a joint extraction framework BSTR-GPRel that integrates a pretrained language model with a global pointer mechanism. The framework achieves the unified modeling of entity relations from ancient texts through a multi-layer architecture. In the input representation layer, BERT sub-word vectors and word embedding features are integrated to analyze the semantic boundaries of complex TCM terms. The semantic encoding layer relies on a BERT-transformer to capture long-distance dependencies and introduces an adversarial training module to enhance the adaptability to ancient text noise. The core entity relation extraction layer uses a three-level global pointer mechanism to ensure the structural integrity of the “subject-relationship-object” triplet. A multi-task joint optimization strategy dynamically balances the loss weights of entity recognition and the prediction of relation heads and tails, specifically addressing the prevalent issue of overlapping relations in traditional Chinese medicine literature. [Result/Conclusion] Experiments were conducted on the corpus of three herbal classics: “Materia Medica for Decoctions”, “Shennong Bencao Jing” and “Materia Medica of South Yunnan”. The results showed that BSTR-GPRel was superior to the comparison model in terms of precision, recall, and F1-score, achieving an F1-score of 0.7854. These findings demonstrate that the proposed model can effectively deal with complex semantic scenarios like overlapping relations. This study provides reliable technical support for the construction of high-quality herbal knowledge graph and helps the implementation of intelligent information service system of traditional Chinese medicine.

Key words： herbal classics; multi-task learning; joint extraction; pre-training model

Cite this article

Shen Wang , Sun Ke , Ma Ruolin , Li Xin , Bian Wei . Research on Joint Extraction of Entity Relations of Herbal Classics Based on BSTR-GPRel[J]. Library and Information Service, 2026 , 70(6) : 30 -39 . DOI: 10.13266/j.issn.0252-3116.2026.06.003

References

[1] 中共中央办公厅、国务院办公厅印发《关于推进新时代古籍工作的意见》[J].中华人民共和国国务院公报, 2022(12):30-33.(The general office of the central committee of the CPC and the general office of the state council print and issue the opinions on advancing the work of ancient classics in the new era[J].Gazette of the state council of the people’s republic of China,2022(12):30-33.)
[2] 曹霞,常存库,裴丽.中医古籍数字化建设及其平台设计和实现[J].中华医学图书情报杂志, 2016, 25(3):45-47,53.(CAO X, CHANG C K, PEI L. Digitalization of ancient books on traditional Chinese medicine and its platform design and implementation[J]. Chinese journal of medical library and information science, 2016, 25(3):45-47, 53.)
[3] 曲毅.基于文本挖掘的《本草集要》语义内容与特征分析[J].中国中医药信息杂志, 2015, 22(12):34-37.(QU Y. Analysis on semantic contents and medication characteristics of Ben Cao Ji Yao based on text mining[J]. Chinese journal of information on traditional Chinese medicine, 2015, 22(12):34-37.)
[4] 李兵,张华敏,符永驰,等.中医古籍本草知识组织方法研究[J].世界科学技术-中医药现代化, 2015, 17(6):1142-1145.(LI B, ZHANG H M, FU Y C, et al. Study on knowledge organization method for ancient herbal books[J]. Modernization of traditional Chinese medicine and materia medica-world science and technology, 2015, 17(6):1142-1145.)
[5] 杨丽丽,李正钧,陈静怡,等.中医古籍芳香类方药外治数据挖掘系统构建及应用[J].中华医学图书情报杂志, 2017,26(8):8-12.(YANG L L, LI Z J, CHEN J Y, et al. Construction of external treatment data mining system for prescriptions of fragrant traditional Chinese medicine covered in ancient books on traditional Chinese medicine[J]. Chinese journal of medical library and information science, 2017, 26(8):8-12.)
[6] YAO L, ZHANG Y, WEI B, et al. A topic modeling approach for traditional Chinese medicine prescriptions[J]. IEEE transactions on knowledge and data engineering, 2018, 30(6):1007-1021.
[7] 杨凤,钟相根,周冉冉,等.基于知识元理论的《伤寒杂病论》标引方案与知识表示研究[J].中国医药导报, 2022,19(30):12-16, 40.(YANG F, ZHONG X G, ZHOU R R, et al.Knowledge indexing scheme and knowledge representation of Treatise on Cold Pathogenic and Miscellaneous Diseases based on knowledge element theory[J]. China medical herald, 2022,19(30):12-16, 40.)
[8] 张卫东,张晓晓.中医古籍数字资源知识组织与可视化研究——以《金匮要略》为例[J].情报科学, 2022, 40(8):107-117.(ZHANG W D, ZHANG X X. Research on knowledge organization and visualization of digital resource based on TCM ancient literature:taking synopsis of the Golden Chamber as an example[J]. Information science, 2022, 40(8):107-117.)
[9] 董美,常志军.一种面向中医领域科技文献的实体关系抽取方法[J].图书情报工作, 2022, 66(18):105-113.(DONG M,CHANG Z J. An entity relation extraction method for scientific and technological documents in the field of traditional Chinese medicine[J]. Library and information service, 2022, 66(18):105-113.)
[10] 何宇浩,李明,罗晓兰,等.基于GPTs的中医知识图谱实体和关系抽取研究[J].上海中医药杂志, 2024, 58(8):1-6.(HE Y H, LI M, LUO X L, et al. Research on entity and relation extraction from traditional Chinese medicine knowledge graphs based on GPTs[J]. Shanghai journal of traditional Chinese medicine, 2024, 58(8):1-6.)
[11] 郭喜跃,何婷婷.信息抽取研究综述[J].计算机科学, 2015,42(2):14-17, 38.(GUO X Y, HE T T. Survey about research on information extraction[J]. Computer science, 2015, 42(2):14-17, 38.)
[12] BLASCHKE C, VALENCIA A. The frame-based module of the SUISEKI information extraction system[J]. IEEE intelligent systems, 2002, 17(2):14-20.
[13] CORNEY D P A, BUXTON B F, LANGDON W B, et al.BioRAT:extracting biological information from full-length papers[J]. Bioinformatics, 2004, 20(17):3206-3213.
[14] DIVOLI A, ATTWOOD T K. BioIE:extracting informative sentences from the biomedical literature[J]. Bioinformatics,2005, 21(9):2138-2139.
[15] 朱姗.基于规则和本体的实体关系抽取系统研究[J].情报杂志, 2010, 29(S2):142-143, 161.(ZHU S. Research on entity relation extraction system based on rules and ontology[J]. Journal of intelligence, 2010, 29(S2):142-143, 161.)
[16] FUNDEL K, KUFFENR R, ZIMMER R. RelEx:relation extraction using dependency parse trees[J]. Bioinformatics, 2007,23(3):365-371.
[17] KAMBHATLA N. Combining lexical, syntactic, and semantic features with maximum entropy models for information extraction[C] //Proceedings of the ACL interactive poster and demonstration sessions. Barcelona:Association for Computational Linguistics, 2004:178-181.
[18] JIANG J, ZHAI C X. A systematic exploration of the feature space for relation extraction[C] //Human Language Technologies2007:the conference of the north American chapter of the association for computational linguistics; proceedings of the main conference. Rochester:Association for Computational Linguistics, 2007:113-120.
[19] MUZAFFAR A W, AZAM F, QAMAR U. A relation extraction framework for biomedical text using hybrid feature set[J].Computational and mathematical methods in medicine, 2015,2015(1):910423.
[20] HINTON G E, SALAKHUTDINOV R R. Reducing the dimensionality of data with neural networks[J]. Science, 2006,313(5786):504-507.
[21] ZENG D, LIU K, LAI S, et al. Relation classification via convolutional deep neural network[C] //Proceedings of COLING2014, the 25th international conference on computational linguistics:technical papers. Dublin:Dublin City University and Association for Computational Linguistics, 2014:2335-2344.
[22] ZHANG Y, YANG J. Chinese NER using lattice LSTM[C] //Proceedings of the 56th annual meeting of the Association for Computational Linguistics. Melbourne:Association for Computational Linguistics, 2018:1554-1564.
[23] WEI Z, SU J, WANG Y, et al. A novel cascade binary tagging framework for relational triple extraction[C] //Proceedings of the58th annual meeting of the association for computational linguistics.Online:Association for Computational Linguistics, 2020:1476-1488.
[24] FU T J, LI P H, MA W Y. Graphrel:Modeling text as relational graphs for joint entity and relation extraction[C] //Proceedings of the 57th annual meeting of the association for computational linguistics. Florence:Association for Computational Linguistics,2019:1409-1418.
[25] ZENG D, LIU K, CHEN Y, et al. Distant supervision for relation extraction via piecewise convolutional neural networks[C] //Proceedings of the 2015 conference on empirical methods in natural language processing. Lisbon:Association for Computational Linguistics, 2015:1753-1762.
[26] PARK C, PARK J, PARK S. AGCN:Attention-based graph convolutional networks for drug-drug interaction extraction[J].Expert systems with applications, 2020, 159:113538.
[27] 李斌,林民,斯日古楞,等.基于提示学习和全局指针网络的中文古籍实体关系联合抽取方法[J].计算机应用, 2025,45(1):75-81.(LI B, LIN M, SIRIGULENG, et al. Joint entityrelation extraction method for ancient Chinese books[J]. Journal of computer applications, 2025, 45(1):75-81.)
[28] 邓亮,齐攀虎,刘振龙,等. BGPNRE:一种基于BERT的全局指针网络实体关系联合抽取方法[J].计算机科学, 2023,50(3):42-48.(DENG L, QI P H, LIU Z L, et al. BGPNRE:A BERT-based global pointer network for named entity-relation joint extraction method[J]. Computer science, 2023, 50(3):42-48.)
[29] SU J, AHMED M, LU Y, et al. Roformer:Enhanced transformer with rotary position embedding[J]. Neurocomputing, 2024, 568:127063.
[30] 段宇锋,柏萍.基于对抗训练和全局指针网络的医疗文本实体关系联合抽取模型[J].情报科学, 2025, 43(3):47-57.(DUAN Y F, BAI P. Joint extraction model for medical text entity and relation based on adversarial training and global pointer networks[J]. Information science, 2025, 43(3):47-57.)
[31] 国家技术监督局.信息处理用现代汉语分词规范:GB/T 13715-1992[S].北京:中国标准出版社, 1992.(National Bureau of Technical Supervision. Contemporary Chinese language word segmentation specification for information processing:GB/T 13715-1992[S]. Beijing:Standards Press of China, 1992.)
[32] SHANG Y M, HUANG H, MAO X. Onerel:Joint entity and relation extraction with one module in one step[C] //Proceedings of the AAAI conference on artificial intelligence. Online:AAAI Press, 2022, 36(10):11285-11293.
[33] ZENG D, ZHANG H, LIU Q. Copymtl:Copy mechanism for joint extraction of entities and relations with multi-task learning[C] //Proceedings of the AAAI conference on artificial intelligence. New York:AAAI Press, 2020, 34(5):9507-9514.
Options
Outlines

/

〈 〉

    Copyrights © LIS Press Co., Ltd. Reproduction is prohibited without permission.