Classification of Electrocardiogram Signals with Deep Belief Networks

Abstract

This paper introduces an electrocardiogram beat classification method based on deep belief networks. This method includes two parts: feature extraction and classification. In the feature extraction part, features are extracted from the original electrocardiogram signal: including features extracted by deep belief networks and timing interval features. Several classifiers are selected to classify the electrocardiogram beat, and nonlinear support vector machine with Gaussian kernel achieves the best classification accuracy, reaching 98.49. Compared with other similar methods on electrocardiogram beat classification, our method can improve the recognition performance of some types of electrocardiogram beats.