基于模糊隶属度与支持向量机心律失常分类模型  

Model of arrhythmia classification based on fuzzy subordination degree and support vector machine

作　　者：杨枢[1] 朱超[2] 

机构地区：[1]蚌埠医学院计算机教研室,安徽蚌埠233030 [2]安徽省蚌埠市第二人民医院心内科,233000

出　　处：《蚌埠医学院学报》2012年第8期985-987,992,共4页Journal of Bengbu Medical College

基　　金：安徽省教育厅高校自然科学研究项目资助(KJ2010B110)

摘　　要：目的:基于心电信号波形特点,运用模糊隶属度与支持向量机技术,探索实现心律失常自动分类的方法。方法:对MIT-BIH心律失常标准数据库的心电信号预处理,识别并定位QRS波;以QRS波为核心,利用心电信号波形相似性进行心电信号聚类;心电信号提取特征参数并模糊化,构建心律失常特征参数集;利用支持向量机技术建立心律失常分类器。结果:通过MIT-BIH心律失常标准数据库检验分类效果,总体准确率达到97.2%。结论:对MIT-BIH心律失常标准数据库的心电信号具有较高的分类准确率和较好的实用性。Objective：To discuss the arrhythmia classification method based on the wave＇s characteristics by using fuzzy subordination degree and support vector machine （SVM） technology. Methods：The electrocardiosignal of MIT-BIH arrhythmia standard database was pre-processed,and QRS waves were identified and located. Electrocardiosignal was clustered by making use of the similarity of electrocardiosignal waves and having QRS waves as the center. The characteristic parameters were abstracted from electrocardiosignal and fuzzified to build arrhythmia characteristic parameter set. The model of arrhythmia classification was established using the technology of SVM. Results：The classification performance which was assessed by the MIT-BIH arrhythmia database reached a total accuracy of 97. 2%. Conclusions： This algorithm has a high accuracy of classification to the electrocardiosignal of MIT - BIH arrhythmia standard database and is quite practicable.

关 键 词：心律失常 动态心电图 模糊隶属度 支持向量机 

分 类 号：R541.7[医药卫生—心血管疾病]