基于特征分类器集成的糖尿病分类方法  

Classification method of diabetes based on integration of characteristic classifier

作　　者：刘丽桑[1] 李强[2] 杨帆[2] 郑哲洲[3] 林雪娟[3] 吴青海[3] 

机构地区：[1]福建工程学院信息科学与工程学院,福州350118 [2]厦门大学自动化系,厦门361005 [3]福建中医药大学中医证研究基地,福州350122

出　　处：《中华中医药杂志》2016年第1期80-83,共4页China Journal of Traditional Chinese Medicine and Pharmacy

基　　金：国家自然科学基金项目(No.81373552);福建省教育厅A类项目(No.JA14212);福建工程学院科研启动项目(No.GY-Z12079)~~

摘　　要：目的:结合医用电子鼻技术,探讨糖尿病患者及其口腔呼气的气味图谱特征。方法:选择180例糖尿病患者和100例健康者,用医用电子鼻采集280例口腔呼气的气味图谱,采用基于数据特征划分的方法,用支持向量机和随机森林集成模型对糖尿病患者进行分类预测。结果:1线性核函数的支持向量机(SVM1)分类结果不是很理想,低于多项式核(SVM2)、径向基函数核(SVM3)和随机森林(RF)3种分类器,说明分类超平面显然是非线性的;2集成分类器对糖尿病患者和健康者的气味图谱特征的识别准确率可达88.04%。结论:基于特征划分的分类器集成方法预测性能明显好于单一分类器,为使用医用电子鼻进行糖尿病诊断分析提供了一种有效手段。Objective： To discuss the proi le features of oral odor of diabetic patients based on medical electronic nose technology. Methods： 180 patients of diabetes and 100 healthy people were selected, and the proi le features of oral odor of 280 volunteers were collected by using medical electronic nose. The classii cation forecasting was carried out on diabetic patients by using support vector machine（SVM） and random forest integration model based on partitioning method of data characteristics. Results： 1The classii cation result of SVM1 was not very good, which was lower than that of SVM2, SVM3 and RF, and the result showed that the classii cation hyperplane is nonlinear. 2The accurate rate of recognition of integrated classii er on diabetic patients and healthy people is 88.04%. Conclusion： The forecasting performance of classii er integration method based on feature division is superior to that of single classii er signii cantly, which provided an ef ective means for the diagnostic analysis of diabetes based on medical electronic nose.

关 键 词：糖尿病 特征分类 模式识别 支持向量机集成 

分 类 号：R587.1[医药卫生—内分泌]