A Critical Evaluation of Machine Learning and Deep Learning Techniques for COVID-19 Prediction

It has been observed that the Corona Virus (COVID-19) patients show signs of abnormality in the chest. It also has symptoms like Palpitation, out of breath, Shakiness, sudation. Recently, several techniques have been Machine Learning (ML) and Deep Learning (DL) techniques, have been proposed for the early detection of this deadly virus. However, the literature still lacks the critical evaluation of the recently proposed techniques to identify the usefulness and applicability of these promising AI-based techniques towards early detections. The work is done in this paper critically reviews and evaluates ML and DL-based techniques. It is a fact that suggesting one technique over other is challenging because of the varying size and diversity of COVID-19 related datasets. However, our review reveals that ResNet-50, VGG16 models of DL, and the Support Vector Machine (SVM), K-Nearest Neighbors (KNN) ML methods are predominantly used in the literature for early detection of Covid using chest X-rays. However, this work identifies the limitation that there are diverse data sets that have been used for validation, and there is a strong need of comparing the ML and DL techniques using the benchmark dataset for the proper and justifiable validation of the prediction results claimed for early prediction of COVID-19 using chest X-ray Images. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.