On the Trade-Off of Accuracy and Computational Complexity for Classifying Normal and Abnormal ECG in Remote CVD Monitoring Systems

Remote cardiovascular disease monitoring systems are characterised from a limited number of available leads and limited processing capabilities. In this paper, we investigate the trade-off between accuracy and computational complexity in order to derive the best strategy for classifying the ECG signal into normal or abnormal in such systems, with the spectral energy contained in the constituent waves of the ECG signal, as the primary feature for classification. Five established classifiers are considered and through exhaustive simulations the maximum accuracy is derived for each classifier. Based on 104 ECG records, we present a systematic analysis of the tradeoff between computational complexity and accuracy, which allow us to deduce the best classification strategy considering only a small number of available leads.