基于文本挖掘技术的《易经》可视化初探

Preliminary Investigation on Visualization of Yi Jing Based on Text Mining Technology

目的基于文本挖掘与可视化技术探索与展现《易经》六十四卦的联系,为中医古籍挖掘提供新思路。方法对《易经》原文进行分词、去停用词等预处理步骤,采用词频统计、Word2Vec词向量模型、词频-逆文件频率文档表示法与关键词抽取、层次聚类分析与相似性网络分析对《易经》文本进行挖掘。结果基于词频统计结果发现,“无咎”在《易经》文本中出现频率最高;基于Word2Vec词向量表示与余弦相似度度量得到,吉和凶有0.734的相似性;层次聚类分析显示,字面含义类似的大过和小过聚在同一大类,互为综卦、字面含义相反的既济和未济聚在不同大类,而字面含义相反的损和益、大有和大过,与互为综卦的泰和否均被聚在同一大类;通过相似性网络分析得到,师和临、损和益、坎和困、噬嗑和萃等10个卦爻对有较强的文本相似性。结论通过文本挖掘技术归纳《易经》的核心思想有无咎、居安思危、物极必反、损益原则,与中医的中庸之道、治未病、阴阳相互转化、损益配伍原则相关。该方法可扩展用于中医古籍的挖掘与可视化研究中。

Objective To explore and illustrate the relationship of sixty-four hexagrams in Yi Jing based on text mining technology and visualization;To provide new ideas for excavating ancient TCM books.Methods After preprocessing steps such as word segmentation and removal of stop words in the original text of Yi Jing,word frequency statistics,Word2Vec word vector model,TF-IDF document representation and keyword extraction,hierarchical clustering analysis and similarity network analysis were used to conduct text mining for Yi Jing.Results Based on the statistical results of word frequency,it was found that“Wujiu”was the word with the highest frequency in the text of Yi Jing.Based on the Word2Vec representation and cosine similarity measure,the similarity between“Ji”and“Xiong”was 0.734.Hierarchical cluster analysis showed that“Daguo”and“Xiaoguo”of similar literal meanings were clustered in the same category,while“Jiji”and“Weiji”of different literal meanings were clustered in different categories.However,“Sun”and“Yi”,“Dayou”and“Daguo”,“Tai”and“Pi”of different literal meanings were clustered in the same category.Through similarity network analysis,it was found that“Shi”and“Lin”hexagrams,“Sun”and“Yi”hexagrams,“Kan”and“Kun”hexagrams,“Shike”and“Cui”hexagrams and other six hexagram pairs had strong correlations with each other.Conclusion Through text mining technology,it is found that the main ideas of Yi Jing are no fault,vigilance in peace time,inevitable reverse after extreme,interaction of profit and loss.These are related to the theory in TCM,including the principle of mediocrity,the prevention treatment of disease,the transformation of yin and yang,and compatibility of medicine based on profit and loss in TCM.This method can be extended to the study of excavation and visualization of ancient TCM books.