Atrial Fibrillation (AF) is the most common arrhythmia. AF has increased peoples health and financial burdens. Patients with AF should be stratified according to a predictive stroke-risk score. According to the complication, risk factor and data of epidemiology, AF is always interesting to be multidisciplinary research topic, one of which is the developing algorithms for auto-detect software. In this study we use analogue recorded of electrocardiography (ECG) data who converted to digital data. Before detecting the presences of AF, we detect R-Peaks of that ECG wave using differential operating method (DOM). Then we analyse the presence of AF by determining irregularities of RR-interval. To detect the occurrence of AF we use two methods, finding anomalies beat around the mean (FAM) and comparison of each other’s beat (CEO). Both of methods are optimized using Particle Swarm Optimization (PSO).The principle of FAM is to look for intervals that have a big margin compared to mean of intervals in a segment. While CEO’s principle is to compared all of intervals in the segment each other, then it find the big different to declare the presence of AF. The role of PSO is to optimize their performance by initializing and evaluating their parameters to create the threshold between normal and AF. We have used this method to test the patient’s data from MIT-BIH. The performance of FAM is presented in accuracy, sensitivity, and specificity of 90.46%, 95.81, and 84.84% respectively. The performance of CEO is presented in accuracy, sensitivity, and specificity of 85.30%, 94.46%, and 77.19% respectively. © Published under licence by IOP Publishing Ltd.
Eliana M., Nuryani N., Satriyo Nugroho A.
Journal of Physics: Conference Series, 10.1088/1742-6596/1153/1/012050