糖尿病证候的数据挖掘方法概述

Summary of data mining methods of traditional Chinese medicine syndromes on diabetes

关键词:  数据挖掘  糖尿病  证候  决策树  关联规则

Key words:  data mining  diabetes  traditional Chinese medicine syndrome  decision tree  correlation analysis

摘要:

[目的] 系统总结运用数据挖掘方法分析糖尿病证候的特征规律。[方法] 归纳近10年采用Logistic回归分析、因子分析、聚类分析、决策树、关联规则、人工神经网络、结构方程模型、贝叶斯网络和支持向量机等数据挖掘方法分析2型糖尿病的文献。[结果] 总结了2型糖尿病的病性要素归纳为气虚、血虚、阴虚和阳虚4种虚证和血热、血瘀、火旺和湿热4种实证，病位要素为肾、肝和脾；主要包括肺热津伤证、胃热炽盛证、气阴两虚证、肾阳虚证、气虚证、痰湿证、血瘀证、肾阴虚证和阴阳两虚证等常见证型，并分析数据挖掘方法的优缺点。[结论] 数据挖掘方法用于糖尿病证候的特征规律的分析是可行的，不仅为糖尿病证候诊断提供客观化依据，还为中医治疗糖尿病提供临床指导。

Abstract:

[Objective] To systematically summarize and analyz ethe characteristics of diabetes syndrome by using data mining method.[Methods] The literature of the characteristicson type 2 diabetes using data mining methods such as Logistic regression analysis,factor analysis,cluster analysis,decision tree,association rule,artificial neural network,structural equation model,Bayesian networks,and support vector machines in the past 10 years was summarized in this paper.[Results] The pathogenic elements of type 2 diabetes are summarized as four types of deficiency syndromes:qi deficiency,blood deficiency,yin deficiency and yang deficiency,and four excess syndromes of blood heat,blood stasis,fire prosperity,and damp-heat. The disease location elements are kidney,liver and spleen. It mainly include the syndrome of lung heat and fluid injury,the syndrome of stomach heat flourishing,the syndrome of qi and yin deficiency,the syndrome of kidney yang deficiency,the syndrome of qi deficiency,the syndrome of phlegm dampness,the syndrome blood stasis,the syndrome of kidney yin deficiency,the syndrome of yin and yang deficiency,and analyzes the advantages and disadvantages of data mining methods.[Conclusion] The data mining method is feasible for the analysis of the characteristics of diabetes syndrome. It not only provides objective basis for the diagnosis of diabetes syndrome,but also provides clinical guidance for TCM treatment of diabetes.