Machine Learning: A Tool to Combat COVID‐19

Abstract

COVID-19 has become a global challenge and is threatening mankind. The global economy is in crisis due to a long tranche of partial to complete lockdown. Forecasting the number of COVID-19 cases is a challenge as cases are both symptomatic as well as asymptomatic, recurrence after recovery is another challenge. Careful data analysis is required to predict and estimate the number of affected cases as well as death ratio. During this pandemic situation, forecasting uncertainty is of utmost importance in decision making. In this chapter, authors have developed a model to predict the COVID-19 confirmed cases. The prediction is based on the data collected in different phases of lockdown in India. In this study, a model is developed using machine learning approaches based on the analysis of data of two Indian states Delhi and Maharashtra where maximum infected cases are found. This study is an attempt to help the decision-makers in better planning and actions. In this study, Neural Network (NN) and M5P model trees are applied to forecast the number of infected cases with each progressive day. Results suggest that the performance of the neural network-based model is slightly better than the M5P model tree in forecasting COVID-19 cases. © 2021 Scrivener Publishing LLC.