ARCHIVES

Abstract: Machine-learning (ML) algorithms have demonstrated their use in predicting perioperative outcomes in order to improve decision-making in future activities. ML models have long been used to identify and prioritise negative threat indicators in a variety of operational sectors. To deal with forecasting issues, a few predictive methods are extensively utilised. The power of the ML model to estimate the number of following patients who will be afflicted by COVID-19, which is currently regarded a severe threat to humanity, is demonstrated in this work. This study used four conventional prediction models to detect risk factors: linear regression (LR), total reduction and operator selection (LASSO), vector support (SVM), and exponential fluctuation (ES). COVID-19 projections have been completed. Each model was given three types of predictions: the number of new viral cases, the number of fatalities, and the number of patients who would be treated in the next 10 days. The outcomes of the study lead to a promising application of these strategies in the current COVID-19 outbreak. The results show that among all the most successful tools for forecasting new confirmed cases, death rate, and recovery rate, ES performs best, followed by LR and LASSO, while SVM incorporates available data sets and performs in all predicted scenarios.