中医证候研究的现代方法学述评(一)——中医证候数据挖掘技术

Modern methodology of TCM syndrome study(I):Data mining technology of TCM syndrome

目的探索中医证候的现代研究方法。方法对近年来的中国中医期刊有关中医证候的数据挖掘技术进行汇总,分析其优势与不足。结果目前用于中医证候研究的数据挖掘方法主要有:关联规则、集对分析、粗糙集理论、聚类分析、人工神经网络、决策树、支持向量机、贝叶斯网络等。结论中医数据具有非线性、模糊性、复杂性、非定量等特征,针对具体的医学数据和不同的挖掘目标往往要将几种方法综合起来应用,以发挥各自的技术优势。

Objective To explore the modern research methods for TCM symptomatology. Method The data mining techniques of TCM symptomatology were summarized from different TCM magazines in recent years, and their advantages and disadvantages were analyzed. Result The result showed that the methods for data mining in TCM symptomatology included association rules, set pair analysis, rough set theory, cluster analysis, artificial neural network, decision tree, support vector machine and Bayes network, etc. Conclusion The TCM data have the characteristics of nonlinearity, indistinction, complicacy and unquantification and so on. These methods should be applied integratedly in accordance with the specific TCM data and different mining aims, and their advantages will be given a full play to.