Wearable ECG signal processing for automated cardiac arrhythmia classification using CFASE‐based feature selection

Abstract

Classification of electrocardiogram (ECG) signals is obligatory for the automatic diagnosis of cardiovascular disease. With the recent advancement of low‐cost wearable ECG device, it becomes more feasible to utilize ECG for cardiac arrhythmia classification in daily life. In this paper, we propose a lightweight approach to classify five types of cardiac arrhythmia, namely, normal beat (N), atrial premature contraction (A), premature ventricular contraction (V), left bundle branch block beat (L), and right bundle branch block beat (R). The combined method of frequency analysis and Shannon entropy is applied to extract appropriate statistical features. Information gain criterion is employed to select features that the results show that 10 highly effective features can obtain performance measures comparable to those obtained by using the complete features. The selected features are then fed to the input of Random Forest, K‐Nearest Neighbour, and J48 for classification. To evaluate classification performance, tenfold cross validation is used to verify the effectiveness of our method. Experimental results show that Random Forest classifier demonstrates significant performance with the highest sensitivity of 98.1%, the specificity of 99.5%, the precision of 98.1%, and the accuracy of 98.08%, outperforming other representative approaches for automated cardiac arrhythmia classification.

An Evolutionary-Neural Mechanism for Arrhythmia Classification With Optimum Features Using Single-Lead Electrocardiogram

“…The problem in this particular case using the accuracy as prediction metric is that normal class has much greater number of samples than arrhythmic samples. Then different types of arrhythmias ventricular, supraventricular, atrial pathologies and their subtypes have 4,8,20,24,26,27,37,53,55,58,68,71,77,80,84,85,86,90,91,94,104,109,118,125,142,143,155,158,166,168,174,176,183,187,190,201,215,224,236,247 different frequency of occurrence some of them rare than others. Accuracy in this case does not put higher importance to the prediction quality of minority classes, which in our case or in the case of disease analysis in general opposes the design objective.…”

“…Recently, researchers have presented different feature reduction methods to reduce input dimensions of ECG signals for neural classifiers. To name a few, Zhang et al [20] extracted statistical features applying the combined method of frequency analysis and Shannon entropy and used information gain criterion to select 10 highly effective features to obtain a good classification on five types of heartbeats. Yildrim et al [21] implemented a convolutional auto-encoder based nonlinear compression structure to reduce the feature size of arrhythmic beats.…”

An Outlier Detection and Feature Ranking based Ensemble Learning for ECG Analysis

“…Filter methods evaluate each feature independently of the classifier, rank the features according to some evaluation criterion and select the best ones [11]. This evaluation can be performed by using entropy for instance [12]. Wrappers methods evaluate the classifier's performance on various subsets of features and select the subset with maximum performance.…”

A Review Study for Electrocardiogram Signal Classification

“…To evaluate classification performance, tenfold cross-validation was used to verify the effectiveness of our method. Experimental results showed that the Random Forest classifier demonstrates significant performance with the SPE of 99.5%, the highest SEN of 98.1%, and the ACC of 98.08%, outperforming other representative approaches for automated cardiac arrhythmia classification [67]. Kora et al (2019) showed that an algorithm to detect atrial fibrillation (AF) in the ECG signal is developed.…”