基于多种信息处理方法的大肠癌中医证候模型构建  

Construction of TCM syndrome model of colorectal cancer based on multiple information processing methods

作　　者：张伟妃 张彦博[2] 张瑞义 李雪 李福凤 ZHANG Wei-fei;ZHANG Yan-bo;ZHANG Rui-yi;LI Xue;LI Fu-feng(Teaching and Experiment Center of Shanghai University of Traditional Chinese Medicine,Shanghai 201203,China;Shuguang Hospital,Shanghai University of Traditional Chinese Medicine,Shanghai 201203,China)

机构地区：[1]上海中医药大学教学实验中心,上海201203 [2]上海中医药大学附属曙光医院,上海201203

出　　处：《中华中医药杂志》2021年第2期1034-1038,共5页China Journal of Traditional Chinese Medicine and Pharmacy

基　　金：国家自然科学基金项目(No.81774205);国家重点研发计划(No.2018YFC1707600)。

摘　　要：目的:基于多种信息处理方法构建大肠癌常见中医复合证候模型。方法:通过流行病学调查方法收集大肠癌患者病例资料1132例,运用聚类分析、BP神经网络、SVM支持向量机和随机森林法对大肠癌患者病例信息进行建模。结果:在BP神经网络、支持向量机和随机森林3种方法构建的大肠癌证候模型中,随机森林效果最好,对各证型的识别率分别为:脾虚气滞(65.1%)、脾肾阳虚(83.3%)、肾精亏虚(92.3%)、湿热蕴结(97.7%)、气血两虚(96.3%)。结论:成功构建大肠癌常见中医复合证候模型,随机森林方法对证型的判断准确率最高,应用随机森林建模法可为中医证候规范化研究提供新的思路和方法。Objective:To construct a model of common TCM complex syndromes of colorectal cancer.Methods:Case information of 1132 patients with colorectal cancer was collected by the epidemiological investigation.Cluster analysis,BP neural network,SVM support vector machine and random forest method were used to model the case information of patients with colorectal cancer.Results:BP neural network,support vector machine and random forest are the three methods to construct the overall recognition rate of colorectal cancer syndrome model.Random forest has the best effect.The recognition rates of each syndrome type are spleen deficiency and qi stagnation(65.1%),spleen and kidney yang deficiency(83.3%),kidney essence deficiency(92.3%),dampness and heat accumulation(97.7%)and qi and blood deficiency(96.3%).Conclusion:The complex syndrome model of TCM for colorectal cancer has been successfully constructed,and the random forest method has the highest accuracy in judging the syndrome types.The application of the Stochastic Forest modeling method can provide new ideas and methods for the standardization of TCM syndromes.

关 键 词：大肠癌 复合证候模型 随机森林 

分 类 号：R273[医药卫生—中西医结合]