基于RBF神经网络的心电分类识别算法研究

Research on ECG Classification Algorithm Based on RBF Neural Network

目前,心律失常是全世界都在研究的热点课题,医生通常都要通过患者的心电图来确定其心脏疾病类型,所以,心电分类识别对此就有重要的意义。该文集中研究了一种RBF神经网络的心电分类识别算法。此算法是利用前向多层神经网络的径向基函数算法（Radial-Basis Function）,即RBF算法,并利用MIT-BIH（美国麻省理工学院提供的研究心律失常的数据库）心电图数据库训练神经网络,使RBF神经网络对未训练过的心电图有较好的分类能力。实验结果表明这种方法用于心电图的分类取得较好的效果。

At present, arrhythmia is a hot research topic all over the world. Doctors usually determine the type of heart disease through the ECG of patients, so ECG classification recognition is of great significance. This paper focuses on an ECG classifica- tion and recognition algorithm Based on RBF neural network. This algorithm is using forward radial basis function algorithm of multi-layer neural network （Radial-Basis Function）, RBF algorithm, and using MIT-BIH （cardiac arrhythmia research provid- ed by the Massachusetts Institute of Technology database） ECG database for training the neural network, the R_BF neural net- work has better classification ability of ECG untrained. The experimental results show that this method is better for ECG classifi-