Comparison of shallow and deep learning methods of ECG signals clas-sification for arrhythmia detection

Abstract

The research aimed to compare the classification performance of arrhythmia classification from the ECG signal dataset from the Massachusetts Institute of Technology–Beth Israel Hospital (MIT-BIH) database. Shallow learning methods that were used in this study are Support Vector Machine,  Naïve Bayes, and Random Forest. 1D Convolutional Neural Network (1D CNN), Long Short Term Memory (LSTM), and Gated Recurrent Unit (GRU) were deep learning methods that were used for the study. The models were tested on a dataset with 140 samples that were grouped into four class labels, and each sample has 2160 features. Those models were tested for classification performance. This research shows Random Forest and 1D CNN have the best performance.