基于摄动思想的中医处方模糊聚类分析

Fuzzy Clustering Analysis of TCM Prescription Based on Perturbation Theory

基于信息技术的中医方剂自动分析,对中医经验的挖掘和传承具有重要意义。本文采用基于摄动思想的模糊聚类方法,对中医方剂进行了聚类分析研究。其中首先对方剂各味药的性味归经毒性进行模糊化处理,然后将基于传递闭包的摄动思想应用到方剂药群的聚类分析。然后以小柴胡汤的聚类分析为例,将此方法与传统的传递闭包矩阵聚类方法进行了比较。实验结果表明,与传统的传递闭包矩阵的模糊聚类方法相比,基于摄动思想模糊的聚类方法效果更好,可以更加客观地反映中医处方组方原则,对于方剂分析具有较好的指导意义。

For mining and inheriting the traditional Chinese medicine （TCM） experience, it is of great importance that information technology is used to analyze the TCM prescriptions automatically. In this paper, the fuzzy clustering method based on perturbation theory was adopted to analyze the TCM prescriptions. First the character, taste, meridian affinity and toxicity of each single medicine in a specific TCM Prescription were processed with fuzzy quantization. Then the fuzzy clustering based on transitive closure matrix of perturbation theory was applied to analyze the prescription composing principles. As an instance, the minor hupleurum decoction was specifically analyzed by such clustering method, and the comparison was made between pure transitive closure method and transitive closure method based on perturbation theory. Experimental results indicated that clustering based on the perturbation theory was capable of revealing the composing principles of TCM prescription more objectively and outperforms the fuzzy clustering based on traditional transitive closure matrix, which had sound guiding significance for prescription analysis.