%0 Journal Article
%T ECG信号自动诊断中回归建模法特征提取的研究
%A 葛丁飞
%J 自动化学报
%P 462-466
%D 2007
%R 10.1360/aas-007-0462
%X ？Thisarticleexplorestheabilityofmultivariateautoregressivemodel(MAR)andscalarARmodeltoextractthefeaturesfromtwo-leadelectrocardiogramsignalsinordertoclassifycertaincardiacarrhythmias.TheclassificationperformanceoffourdifferentECGfeaturesetsbasedonthemodelcoefficientsareshown.Thedataintheanalysisincludingnormalsinusrhythm,atriaprematurecontraction,prematureventricularcontraction,ventriculartachycardia,ventricularfibrillationandsuperventriculartachycardiaisobtainedfromtheMIT-BIHdatabase.Theclassificationisperformedusingaquadraticdiscriminantfunction.TheresultsshowtheMARcoefficientsproducethebestresultsamongthefourECGrepresentationsandtheMARmodelingisausefulclassificationanddiagnosistool.
%K Autoregressivemodel
%K ECGfeatures
%K classification
%K automaticdiagnosis
%U http://www.aas.net.cn/CN/abstract/abstract14291.shtml