An ECG-based feature selection and heartbeat classification model using a hybrid heuristic algorithm

This paper proposes a hybrid model to classify cardiac arrhythmias and select their features in an optimal way. In the proposed model, the Genetic Algorithm was used to optimally select the features, and the Decision Tree with the C4.5 algorithm was applied to the extracted features to classify and train the model. The proposed approach was used to classify data into normal and abnormal classes as well as a 16-class collection of arrhythmias. To evaluate the performance of the proposed model compared with similar methods, we used the UCI arrhythmia dataset along with accuracy, sensitivity, specificity, and average Sen-Spec metrics. The efficiency of the proposed method in both two-class and 16-class modes significantly improved the accuracy, sensitivity, the average of sensitivity and specificity parameters compared to similar methods. Our approach obtained values of 86.96%, 88.88%, and 86.55% for the two-class mode and 78.76%, 76.36%, and 78.69% for the 16-class mode classification in terms of accuracy, sensitivity, and the average Sen-Spec metrics respectively. The above-mentioned values are reported as the highest for the UCI arrhythmia dataset.