基于CiteSpace软件中医数据挖掘文献的可视化分析研究

目的通过对有关中医数据挖掘方面的文献进行计量、可视化分析,从而探究其研究历史、现状、热点以及发展趋势。方法在CNKI检索自建库以来有关中医学数据挖掘方面的文献,应用Excel对纳入的文献进行初步统计,应用CiteSpace对文献中的作者、机构、关键词进行共现分析,绘制相关知识图谱,使用模块值和平均轮廓值评价视图的结构和清晰度,通过突现度和中介中心性发现不同聚类中的重要机构、作者以及关键词,对各主要关键词进行聚类分析并采用对数似然比方法对聚类进行标记,使用时间线视图展示相关聚类、聚类之间的相互影响以及聚类中重要关键词的历史跨度。结果共获得文献3019篇,发文量最大的三家机构是北京中医药大学、中国中医科学院、广州中医药大学,起“桥梁”作用的机构有14家;核心作者有59位,共发文955篇,占总发文量33%,其中吴嘉瑞、任玉兰、张冰、周雪忠等学者在国内有较大影响力;在中医数据挖掘主题下共有关键词264个,其中每5年出现频率≥21次有45个,起“桥梁”作用的关键词有25个,形成有意义的聚类12个。结论中医证候分布、用药规律、选穴规律、传承名医经验思想依然是中医数据挖掘方面的热点;中医数据挖掘技术的热点主要集中在聚类分析、关联规则、logistic回归,其他技术例如因子分析、复杂网络、决策树、主成分分析、频数分析、贝叶斯网络、神经网络等应用方面仍较薄弱。

Visualization Analysis of Literature of TCM Data Mining Based on CiteSpace Software

Objective To explore the histroy,present situation,development trend and the research hotspots of the field of data mining of traditional Chinese medicine(TCM),relevant literature was visualized and analyzed.Methods In CNKI,literature on data mining of traditional Chinese medicine had been searched since the establishment of the database.Excel was used to make preliminary statistics of the included documents.CiteSpace was used to analyze the authors,organizations and keywords in the literature and draw the relevant knowledge map.Module values and average contour values were used to evaluate the structure and clarity of views.The important organizations,authors and key words in different clusters were found by using the degree of emergence and intermediary centrality.The main keywords were analyzed and labeled by log likelihood ratio.The time line view was used to show the correlation clustering,the interaction between clusters and the historical span of key words in clustering.Results A total of 3019 papers were obtained.The three institutions with the largest number of articles published were Beijing University of Chinese Medicine,China Academy of Chinese Medical Sciences and Guangzhou University of Chinese Medicine.There were 14 institutions that play a role of“bridge”.There were 59 core authors,with 955 papers distributed,accounting for 33%of the total volume,among which Wu Jia-rui,Ren Yu-lan,Zhang Bing,Zhou Xue-zhong and other scholars had great influence in China.There were 264 key words under the theme of TCM data mining,and 45 of them are frequency greater than or equal to 21 times every five years.There were 25 key words to play the role of"bridge".Finally,12 meaningful clusters were formed.Conclusions The distribution of TCM syndromes,the regularity of drug use,the law of selecting points,and the idea of inheriting famous doctors are still the focus of data mining in TCM.The focus of TCM data mining technology focuses on cluster analysis,association rules,logistic regression,but the application aspects of other technologies such as factor analysis,complex network,decision tree,principal component analysis,frequency analysis,Bayesian network,neural network and other applications are still weak. Using the new platform andnew method to study TCM may be the trend of data mining of TCM in future.