基于LDA模型的中医药人工智能领域主题演化分析＊

目的 梳理总结中医药人工智能领域近40年的研究情况,分析该领域研究主题演化过程,发现热点主题和研究前沿。方法 本研究检索了CNKI和WOS中中医药人工智能领域的文献,根据生命周期理论划分不同时间片,将文献摘要作为研究对象,采用LDA主题模型挖掘文本主题,分析主题的演化过程,识别热点主题。结果 纳入中英文文献2750篇,根据生命周期理论划分中医药人工智能领域发展阶段,分别建立全局和各阶段的LDA模型,挖掘出全局热点主题有中医专家系统、中医辅助诊断系统、中医药数据挖掘等,发现主题演化路径并进行可视化展示。结论 中医药与人工智能主要的结合点在中医智能系统、中医药知识组织和中医智能诊疗机器人这3方面,这3个方向的研究均离不开神经网络技术的支持,同时也结合了关联规则发现、复杂网络研究等方法。

Analysis of topic evolution in the field of artificial intelligence in traditional Chinese medicine based on LDA model

Objective Summarize the research situation in the field of artificial intelligence of traditional Chinese medicine in the past 40 years, analyze the evolution process of research topics in the field, and discover hot topics and research frontiers.Method We searched papers in the field of artificial intelligence of traditional Chinese medicine in CNKI and WOS, divided time slices according to the life cycle theory, mine text topics using LDA model and analyzed the evolution process of the topics to identify hot topics.Result A total of 1995 Chinese literatures are included in this paper. We divided the development stages of the field of artificial intelligence in Chinese medicine according to the life cycle theory, established the LDA models of the overall and each stage, and mine the global hot topics such as Chinese medicine expert system, Chinese medicine auxiliary diagnosis system, Chinese medicine data mining, etc. We also discovered and visualized evolution path of the topics in artificial intelligence of TCM.Conclusion The integration parts of Chinese medicine and artificial intelligence are in the three aspects of Chinese medicine intelligent system, Chinese knowledge organization and Chinese medicine intelligent diagnosis and treatment robot. Research in these three directions were main supported by neural networks, association rules and complex networks.