基于径向基函数神经网络的心脏早搏分类诊断研究

早搏是常见的心律失常，根据发生部位分为窦性早搏、房性早搏、交界性早搏和室性早搏。心肌细胞特有的电生理特性和心律失常中常有的一些心电现象，使一部分早搏的心电图失去其固有的特征，临床上通过心电图对其分类诊断存在一定的不确定性。基于Matlab平台，采用径向基神经网络方法，对所搜集的82个早搏分类确诊病例进行研究，建立了心脏早搏分类的辅助诊断模型，其准确率达到96％。为临床心脏早搏分类诊断提供了一种新的方法。

Diagnostic study on the classification of premature beat based on RBF artificial neural network

Premature beat is a common kind of cardiac arrhythmia. It can be classified into four types according to its origins: sinus, atrial, junctional and ventricular contractions. Some premature beats lose their characteristics in electrocardiogram due to electrophysiological characteristics of cardiac muscle cells and some electrocardiac phenomena in cardiac arrhythmia. These factors make it difficult in diagnosis of premature beat by electrocardiogram in clinic. This study investigated 82 electrocardiograms of premature beat by radial basis function artificial neural network based on Matlab. A model, which provides a new method for the diagnosis of premature beat, has been established and its diagnostic accuracy was 96%.