<![CDATA[The heart, a vital organ responsible for pumping oxygenated blood through blood vessels, is susceptible to disturbances in heart rate that can have adverse effects. According to data from the World Health Organization (WHO) since 2000, this disease has experienced the most significant increase in fatalities, rising from over 2 million to 8.9 million deaths. The prediction of Sudden Cardiac Death (SCD) continues to gain attention as a promising approach to saving millions of lives threatened by the occurrence of the disease. In this study, we propose the utilization of Particle Swarm Optimization (PSO) as a feature selection method to train the Support Vector Machine (SVM) and Logistic Regression. By employing the proposed algorithm, SCD can be predicted up to 30 minutes before the onset with an accuracy of 92.5%, by using PSO and SVM. Features are extracted from Heart Rate Variability (HRV) analysis and Discrete Wavelet Transform (DWT) obtained from ECG records of MIT-BIH normal sinus rhythm database & MIT-BIH Sudden Cardiac Death Holter database dataset. This paper also compares feature selection algorithm of PSO and Analysis of Variance (ANOVA) and found that PSO is better in accuracy, recall, and F1-score.]]>
Copyrights © 2024
