基于改进LS-TWIN-SVM的心律不齐异常辅助诊断  

Auxiliary Diagnosisof Arrhythmia Based on Improved LS-TWIN-SVM

作　　者：杨青峰[1,2] 夏芳 高海燕 冯晓菊[1] 程顺达 YANG Qingfeng;XIA Fang;GAO Haiyan;FENG Xiaoju;CHENG Shunda(Hebei Provincial Hospital of traditional Chinese medicine,Shijiazhuang 050011,China;Hebei University of Economics and Trade,Shijiazhuang 050011,China;Hebei Traditional Chinese Medicine Development Center,Shijiazhuang 050011,China)

机构地区：[1]河北省中医院,河北石家庄050011 [2]河北经贸大学,河北石家庄050011 [3]河北省中医药发展中心,河北石家庄050011

出　　处：《微型电脑应用》2023年第1期101-103,107,共4页Microcomputer Applications

摘　　要：针对社区医疗技术限制问题,结合采集到的患者心电信息,提出了一种改进最小二乘双子支持向量机的信号诊断方法,并在MIT-BIH数据库中对算法性能和自动诊断识别进行仿真实验。结果表明,基于有向无环图的LS-TWIN-SVM分类算法的计算复杂度最低,且其分类及诊断正确率达到99.32%。说明该算法对深入研究生物信号心电信号识别,实现大数据辅助诊断心律不齐异常症状具有实际意义和应用价值。To solve the problem that traditional biological signal recognition cannot realize effective automatic diagnosis due to the limited medical resources in China, a least squares twin support vector machine(LS-TWIN-SVM) classification algorithm based on least square method and support vector machine was established, and the classification strategy of the algorithm was improved by using directed acyclic graph. The performance of the algorithm and automatic diagnosis and recognition were simulated in MIT-BIH database. The results show that the LS-TWIN-SVM classification algorithm based on directed acyclic graph has low computational complexity, can obtain classification results quickly, and its classification and diagnosis accuracy rate is high, it reaches 99.32%. It shows that the algorithm has practical significance and application value for the in-depth study of biological signal ECG signal recognition and the realization of big data aided diagnosis of arrhythmia symptoms, which lays a theoretical foundation for the establishment of community-oriented big data platform for family doctor contract service.

关 键 词：心电信号 最小二乘双子支持向量机 心律不齐 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]