Performance Comparison of Various Optimization Techniques for ECG Signals

Feature extraction from the ECG (Electrocardiogram) data is a most important phase for the identification of heart diseases. As a result, the value of the ECG categorization approach in the present medical world is quite high. Many machine learning programmes are available that may be used to classify and analyse ECG data. The primary flaw in most ECG classification methods is that they have non-optimized capabilities for unstructured function learning frameworks. The problem originates from the lack of methods that might accurately handle the task of ECG categorization. The use of optimisation algorithms coupled with the deep learning idea is the apparent answer among the different strategies that are now accessible. The crucial part is figuring out the right R-peaks and RR intervals. In the suggested study, performance metrics of various meta heuristic optimisation techniques i.e. Cuckoo search along with Support Vector Machine(SVM), Sine Cosine Algorithm (SCA) and Linear Adaptive Sine Cosine Algorithm(LA-SCA) are compared. To assess and train the system, signal data is used from the MIT-BIH arrhythmia database. The main goal of this paper is to obtain information about various arrhythmia classes. Considering accuracy and precision as performance parameters, the comparison of above said algorithms is presented.