ABSTRACT
The outbreak of the novel coronavirus pandemic (COVID-19) caused severe threats to humankind across the globe. The COVID-19 virus fits to the large family of virus that stimulate illness that may range from common flu to severe diseases like Middle East respiratory syndrome (MERS-CoV) and severe acute respiratory syndrome (SARS-CoV). Hence, the virus affects the mankind variably ranging from mild to moderate and sometimes very severe leading to mortality. The virus is contagious, and necessary prevention and protection mechanism protocols have been strictly adhered by the public to prevent the community spread which has not succeeded. Researchers have investigated that the spread of the severity of spread could be achieved through herd immunity. This pandemic outbreak has affected the globe to a great extent;hence, effective mechanisms are under investigation to diagnose the disease at the initial stages to prevent spread. Machine learning (ML), a subset of artificial intelligence (AI), provides models that have the ability to inevitably learn and evolve over experience. The ML algorithms are used in diverse applications, its contribution to medical management especially in preventive medicines, medicinal chemistry, imaging, and genetic medicines are inevitable. The sovereign intelligence and capability of ML algorithms make it manifest to use it in the COVID-19-based research. This chapter focusses on usage of ML algorithms to detect the severity of COVID-19 virus in human kind. The model investigates on predicting the severity of risk, risk of infection, and who is at risk of developing a severe case. The ML algorithms described in this chapter aims at identifying the presence of the disease in a patient. This work analyzes the foreseen of the diseased people from people with minor indications built on 111 impute relating to medical and the clinical examination facts. The diagnostic knowledge entailed attributes, such as age group, gender, body temperature, respiratory proportion, heart rate, and BP. The blood/urine examination data contain information related to various categories of blood examination values and urine examination values. Numerous ML models, such as Naive Bayes, SVM, artificial neural network, k-nearest neighbor (kNN), convolutional neural network (CNN), logistic regression, and decision tree were used in prediction and severity analysis. The experimentation may investigate the effectiveness of the above algorithms for detecting the patients infected with the virus as well as the severity level. Further, this research could be enhanced to provide treatment recommendations in the future. © 2022 Scrivener Publishing LLC.