基于自然语言处理智能技术的中医术语研究文献计量分析

Bibliometric analysis of traditional Chinese medicine terminology research based on natural language processing technologies

目的对国内外近20年来发表的涉及自然语言处理(NLP)智能技术应用于中医术语识别或标注方面的文献进行计量分析与评价,探讨NLP智能技术在中医术语标准研究中的应用和发展前景。方法检索收集2003年1月至2023年10月期间,中国知网、维普中文科技期刊数据库、万方数据知识服务平台、中国生物医学文献服务系统及Web of Science等中英文数据库中的相关文献。采用Excel vba、Gephi、PyCharm等数据处理和统计分析工具,应用频数统计、Apriori关联分析、词云统计等文献计量学方法,对相关研究热点进行可视化分析。结果①经筛选,符合研究标准的文献共442篇,其中中文文献320篇、英文文献122篇。②2016年以后,相关发文量呈现持续增长的趋势。③发文国家主要集中在中国。④中文文献中硕博士学位论文比重较大,其中发文量最高的是北京交通大学。⑤中文文献发文机构以中国中医科学院发文量最高;英文文献发文机构以北京科技大学发文量最高;中医机构与计算机相关机构合作频繁。⑥基于BERT的命名实体识别算法在中医术语研究中的应用效果最为显著。⑦中医文献类的数据占比较大。结论基于NLP智能技术的中医术语标准化研究仍处于探索阶段,现有研究表现出技术应用的多样性,但缺乏系统性。鉴于NLP智能技术在中医术语识别和标注方面的潜力,未来研究需进一步加强,以期实现中医术语标准研究的系统化、智能化与广泛应用。

Objective To conduct a bibliometric analysis and evaluation of domestic and international literature published in the last 20 years on the application of natural language processing(NLP)technologies in the identification and labelling of traditional Chinese medicine(TCM)terminology,and explore the application and future development prospects of NLP technologies in the standardization research of TCM terminology.Methods Literature from January 2003 to October 2023 was retrieved from databases including China National Knowledge Infrastructure(CNKI),VIP Chinese Science and Technology Journal Database(VIP),Wanfang Data Knowledge Service Platform(Wanfang),China Biomedical Literature Service System(SinoMed),and Web of Science.Data processing and statistical analysis tools such as Excel vba,Gephi,and PyCharm were used,employing bibliometric methods like frequency statistics,Apriori association analysis,and word cloud statistics to visually analyze the research hotspots.Results①After screening,442 papers met the research criteria,comprising 320 in Chinese and 122 in English.②Publications showed a consistent growth trend from 2016 to 2023.③The majority of publications were from China.④Among the collected Chinese literature,a significant proportion of publications were master's and doctoral theses,with the highest number of publications from Beijing Jiaotong University.⑤The China Academy of Chinese Medical Sciences had the highest publication count in Chinese,while University of Science and Technology Beijing led in English publications.There was frequent collaboration between TCM institutions and computer science institutions.⑥The application of named entity recognition(NER)with BERT showed the most significant effects in TCM terminology research.⑦Data related to TCM literature accounted for a substantial proportion.Conclusions NLP-based research on the standardization of TCM terminology is still in the exploratory stage.Previous studies have shown the diversity of technology applications but lack systematization.Considering the potential of NLP technologies in TCM terminology recognition and labelling,further research is required to achieve systematic and intelligent TCM terminology standardization research and its widespread application.