Deer Hunting Optimization with Deep Neural Network for Automated Arrhythmia diagnosis in ECG Signals

The irregular electrical pattern of the heart is referred as Cardiac arrhythmia, which can be analyzed by the attributes of electrocardiogram (ECG). In accordance with recent survey, 60 million people in worldwide were at danger for Cardiac arrhythmia diseases. Owing to the convenience and noninvasiveness, the ECG waveform is generally considered to determine cardiac arrhythmias disorders. In Cardiology field, immediate attention is required for an effective automatic detection model to detect Cardiac arrhythmia. The aim of this research work is to create an automated ECG arrhythmia detection system for accurate diagnosis. In the first step, the ECG signal is preprocessed to eliminate the unwanted artifacts present in the signal. In the second step Welch method is applied to evaluate the power spectrum of the ECG signal. The ECG signals statistical characteristics, such average peak distance, maximum peak signal strength, and mean value of peak distance, energy and Spectral Skewness are derived from the signal’s spectrum. In the final step, deer hunting optimization (DHO) based neural network classifier is used to classify normal and abnormal ECG signals based on the extracted features. The DHO based NN demonstrated better outcomes with maximal precision of 92.4%.