An Effective Breathing Pattering Classification using Opposition Learning based Ebola Optimization Algorithm and GRU With Clipping Activation Function

Many respiratory diseases, such as sleep apnea, asthma, Chronic Obstructive Pulmonary Disease (COPD), and COVID-19 are manifest through abnormal breathing patterns. A Deep Learning (DL) model automatically and accurately classify these patterns, enabling early detection and intervention before the condition worsens. This research introduced the DL model for classify the breathing pattern using collected dataset. The Gated Recurrent Unit with Clipping Activation function (GRUCA) is used to classify the patterns. Initially the dataset is pre-processed for effective classification and extracting features by time domain, frequency domain, and time frequency domain analysis. The novelty of the work is developing Opposition Learning Based Ebola Optimization Algorithm (OB-EOA) for selecting the best combination of the feature subsets. The opposition strategy enhancing the search ability of the optimization algorithm. The proposed OB-EOA-GRUCA model achieves the promising result with accuracy of 99.57% when compare with the other exiting model such as Stacking Convolutional Neural Network (CNN).