Automatic Identification Prevention And Treatment Of Covid 19 Using Deep Learning In Iot Healthcare System

Abstract

COVID disease still spreads all over the world through changing its structure and developing a new variant. Acute respiratory syndrome with the corona virus, which affects the lungs of people and causes pneumonia, is the source of COVID-19. Breathing becomes difficult due to inflammation and fluid buildup in the lungs caused by pneumonia. Three novel COVID-19 detection methods are developed to overcome the above-mentioned issues. In the first method, using numerous ML approaches with reptile search optimization-based feature selection is developed to predict the exact condition of a person. The raw dataset is pre-processed and extract the features, after that choosing the exact attributes to given the ML classifiers based on KNN, NB, SVM, and XGB for predicting the disease. In second approach using optimized k-means clustering based hybrid VGG19-support vector machine to detect the COVID-19 in IoT devices. In the end, the hybrid VGG19-SVM divides three categories as determined by the CXR. Finally, in third work, a watershed segmentation with hybrid CNN-LSTM model to predict COVID 19 in CT image.. The three suggested models are used with Python software to investigate the implementation of suggested model to identification and prevention COVID 19. The first approach of ML based RSA model provide accuracy of 98% in SVM, 92% in XGBoost, 88% in KNN, 85% in NB and 81% in RF. In second approach, the hybrid VGG19-SVM model offer 98% accuracy, 97% recall, and 0.008% FPR. Also, in third work, the hybrid CNN-LSTM prediction model offer 0.93 sensitivity, 0.97 accuracy, 0.92 F1_score, and 0.08 FPR. To verify the performance of the suggested model, the results were contrasted with those of other current techniques. newline