Comparison of Soft Computing and Optimization Techniques in Classification of ECG Signal

Electrocardiogram (ECG) is a visual representation of the heartbeat that can be used to detect cardiac problems. It helps in detection of a normal or abnormal state of heart diseases. Therefore, it is difficult to detect the cardiological status with naked eyes. So, features extraction from ECG signal is crucial to recognise heart disorders. After selecting significant features, classification can be done by Machine Learning (ML), and Deep Learning (DL). Most of the methods utilised to classify the electrocardiogram are based on 1-D electrocardiogram data. These methods focus on extracting the attributes wavelength and time of each waveform as input but but their behavior remains different during classification. Various ECG construal algorithms based on signal processing approaches have been planned in recent years. Few studies show how optimisation techniques are helpful for feature selection and classification with ML and DL. This work compares the studies based on ML and DL. It also depicts how optimisation methods increase the accuracy, sensitivity, and specificity of data.