<![CDATA[Arrhythmia is one of the most dangerous heart rhythm disorders, especially for the elderly, due to degenerative changes in cardiac structure and function. This study aims to develop an electrocardiogram (ECG)-based arrhythmia classification model for the elderly using the Support Vector Machine (SVM) algorithm. Data were collected from three nursing homes with a total of 184 subjects aged 50–75 years using the Smart Holter ECG 5-lead device. The research stages included ECG signal acquisition, signal preprocessing (baseline correction and Butterworth filter), physiological feature extraction (PR, QRS, QT, RR intervals, ST segment, heart rate, R/S ratio), and data labeling by cardiologists. The model was trained and tested using a hold-out approach with an 80:20 ratio and class stratification. Evaluation results showed high performance with 96.36% accuracy on the training set and 94.57% accuracy on the testing set. The Area Under Curve (AUC) reached 0.99 in micro-average and 0.98–1.00 for each class. This research confirms that SVM is effective for arrhythmia classification in the elderly and has potential as an accurate and efficient diagnostic tool]]>
Copyrights © 2025
