Using feed forward neural network for electrocardiogram signal analysis in chaotic domain

A feed forward neural network for classification of the Electrocardiogram (ECG) beats is employed in this paper. The classification is performed based on a feature extraction scheme. Six groups of ECG beats (MIT-BIH Normal Sinus rhythm, BIDMC congestive heart failure, CU ventricular tachyarrhythmia, MIT-BIH atrial fibrillation, MIT-BIH Malignant Ventricular Arrhythmia and MIT-BIH Supraventricular Arrhythmia) are characterized by the six chaotic parameters including the largest Lyapunov exponent and average of the Lyapunov spectrum (related to the chaoticity of the signal), time lag and embedding dimension (related to the phase space reconstruction) and correlation dimension and approximate entropy of the signal (related to the complexity of the signal). Nine different structures of the feed forward neural network (in terms of the hidden neurons and the number of learning epochs) were employed to perform the ECG beats classification based on all extracted features for three lengths of the signals. The effects of signals length and FFNN structural parameters are studied in the experiments. It is shown that a minimum length is required for valid signal characterization and also it is shown that how accurate the trained feed forward neural network can discriminate the signals of ECG beats for different arrhythmia behaviors. Some conclusions are made and they are discussed at the end of the paper.