基于线性与非线性特征融合的J波自动识别  

Automatic recognition of J wave based on combination of linear and nonlinear features

作　　者：毋凡铭 李灯熬 赵菊敏 Wu Fanming;Li Deng’ao;Zhao Jumin(College of Information & Computer,Taiyuan University of Technology,Jinzhong Shanxi 030600,China)

机构地区：[1]太原理工大学信息与计算机学院

出　　处：《计算机应用研究》2019年第8期2548-2551,共4页Application Research of Computers

基　　金：国家自然科学基金面上项目(61371062,61772358);山西省国际科技合作项目(201603D421014)

摘　　要：临床研究表明,J波可作为一些心脏疾病的高危预警指标。针对当前医生诊断J波仅通过经验进行识别,易造成误诊的问题,从信号处理角度,提出了一种J波自动识别方法。通过提取心电数据极点对称模态分解后的能量特征与高阶累计量特征,融合线性与非线性特征,并采用主成分分析进行特征降维,最后利用人工蜂群算法优化后的支持向量机进行分类,实现J波的自动识别。对比实验结果,提出的方法平均准确率为97.3%,可有效地识别J波。Clinical studies have shown that J wave can be used as a high risk early warning index of some heart diseases.In view of the shortcomings of J wave diagnosis only by the clinicians’ experiences,which can easily lead to misdiagnosis.From the perspective of signal processing,this paper proposed a J wave automatic identification method.This method extracted the energy features of electrocardiogram data after the extreme-point symmetric mode decomposition and higher order statistics,combined linear and nonlinear features,and adopted the principal component analysis to reduce the dimension of the features.Finally,it used support vector machine optimized by artificial bee colony algorithm to realize automatic identification of J wave.The experimental results show that the average accuracy rate of identifying J wave is 97.3%,which can effectively identify the J wave.

关 键 词：J波 极点对称模态分解 高阶累计量 人工蜂群算法 支持向量机 

分 类 号：TP302.1[自动化与计算机技术—计算机系统结构]