An Efficient Prediction of Obstructive Sleep Apnea Using Hybrid Convolutional Neural Network

Abstract

Obstructive sleep apnea (OSA) represents a severe sleep disorder, which exhibited during the interruption of the breathing while slee**. It causes as a result of inadequate supply of oxygen to both the brain and the physical body. In this research work, Convolutional Neural Network (CNN) has been used for feature extraction and selection and the Artificial Neural Network (ANN) has been used to predict sleep apnea. A flattened layer, fully connected networks, and feature extraction layers make up the suggested CNN model. LSTM model is used for classification, composed of Softmax classification layer to identify disease classification and to compute its association with cardiovascular diseases. Experimental analysis of the proposed framework toward the prediction of obstructive sleep apnea classification has been done using the Physionet apnea dataset. This dataset aims to evaluate the effectiveness of the proposed representative framework against the conventional approaches to analyze in various deep learning architectures. The proposed framework achieves accuracy of 99% on optimal feature classification for disease prediction against machine-based existing classification approaches.