Comparison Analysis for Life-Threatening Arrhythmia Classification from ECG Data Using Machine Learning and Deep Learning Methods

Arrhythmia, characterized by abnormal heart rhythms, poses significant risks to patients with cardiovascular diseases. Timely detection of life-threatening arrhythmias is essential for the effective monitoring and management of such conditions. Numerous studies have focused on the detection and classification of electrocardiogram (ECG) signals to enable prompt and appropriate resuscitation interventions. This study presents a comparative analysis of various arrhythmia classes within the MIT-BIH Malignant Ventricular Ectopy Database (MVED). This study employs Fast Fourier Transform (FFT) and Daniell's Periodogram Estimate as feature extraction techniques. To classify the arrhythmias, several machine learning algorithms were implemented, including Convolutional Neural Networks (CNN), K-Nearest Neighbors (KNN), and Support Vector Machines (SVM). Among these, the CNN model demonstrated superior performance, achieving an accuracy of 89% in multi-class classification tasks.