Pro-detection of atrial fibrillation with ECG parameters mining technique

Abstract

Reliable detection of atrial fibrillation (AF) in ECG monitoring systems is significant for early treatment and health risks reduction. Various ECG mining and analysis efforts have addressed a wide variety of clinical and technical issues. However, there is still scope for improvement mostly in the number and the types of ECG parameters necessity to detect AF arrhythmia with high quality that encounter a massive number of challenges in relation to computational efforts and time consuming. In this paper, we proposed a technique that caters these limitations. It select features related to the ECG parameters, so as to design a unique feature set that could be employed to describe AF in very sensitive manner. The performance of our proposed technique showed a sensitivity of 95% and a specificity of 99.6%, and overall accuracy of 99.2%. © 2012 Springer-Verlag.