机器学习方法在中医证候分析中的应用

Application of Machine Learning Method in TCM Syndrome Analysis

传统的中医辨证诊疗主要基于"望、闻、问、切"得到的四诊信息,由于掺杂过多的医师主观因素,即使对同一个病人的辨证结果也可能不尽相同,因此如何建立一个科学而规范化的中医证候的量化标准是一个值得研究的课题。本文将机器学习中的层次聚类和因子分析方法应用于中医证候量化分析,通过对采集到的1499例典型高血压病例的处理与分析,实验结果表明,机器学习方法可以有效地挖掘中医证候中隐藏的信息,为中医辅助诊疗提供重要的途径。

The traditional diagnosis and treatment of TCM syndrome differentiation is mainly based on the information of"looking,smelling,asking and cutting".Because of too many subjective factors of doctors,even for the same patient,the results of syndrome differentiation may be different,so how to establish a scientific and standardized quantitative standard of TCM syndrome is a sub⁃ject worthy of study.In this paper,the hierarchical clustering and factor analysis methods in machine learning are applied to the quantitative analysis of TCM syndromes.Through the treatment and analysis of 1499 cases of typical hypertension collected,the ex⁃perimental results show that machine learning method can effectively mine the hidden information in TCM syndromes,and provide an important way for TCM assisted diagnosis and treatment.