摘要
为了给医生在心血管疾病诊断方面提供更精确的参考依据,提高心血管疾病诊断效率,提出了一种基于PCA-SVM模式分类的心电信号分析方法。通过对麻省理工心率失常数据库中8类心搏心电数据分别运用支持向量机以及PCA-SVM模式分类方法进行分析处理,比较最终的分类准确率。发现当支持向量机选择线性核函数时,SVM的分类准确率为97.812 5%,PCA-SVM的分类准确率为99.0625%,PCA-SVM相对于SVM的分类准确率更高,能够满足心血管疾病临床诊断需求。
To provide more accurate reference for doctors in the diagnosis of cardiovascular disease and improve the diagnostic efficiency of cardiovascular disease. Proposing a method of ECG signal analysis based on PCA-SVM mode classification.By analyzing the 8 kinds of cardiac electrocardiogram data in the heart failure database of MIT, we used SVM and PCA-SVM pattern classification method to analyze and compare the final classification accuracy. It is found that when support vector machine selects linear kernel function, the classification accuracy of SVM is 97.812 5%, the classification accuracy of PCA- SVM is 99.0625% , classification accuracy of PCA-SVM is higher than SVM , and it can meet the clinical diagnosis needs of cardiovascular disease.
作者
卞水荣
顾媛媛
赵强
BIAN Shui-rong;GU Yuan-yuan;ZHAO Qiang(School of Medical Information,Xuzhou Medical College,Xuzhou 221000,China)
出处
《电子设计工程》
2018年第20期37-41,共5页
Electronic Design Engineering
基金
国家重点研发计划项目(2017YFC0804401)
江苏省产学研联合创新项目(BY2014033)
徐州医科大学校课题(2015SK01)
作者简介
卞水荣（1991-）,男,江苏南通人,硕士.研究方向：远程医疗、移动医疗.
