Statistical Analysis and Prediction of COVID-19 outbreak in India using Machine Learning

Abstract

Coronavirus disease globally known as COVID-19 is triggered by SARS-COV2. It is the predominant cause of an extremely dangerous disease that has bothered global health security. It is proposed that COVID-19 might be zoonotically based on the high number of people exposed in Wuhan City, China, to the wet animal market[1]. COVID-19 is a severe acute respiratory disease, transmitted by respiratory secretions and communication paths, as of WHO reports. The disease is spreading throughout the world at a faster pace. The first instance of COVID-19 was firstly discovered and found in Wuhan, Hubei Province, China in December 2019[1]. This paper analyses the outbreak of this disease until June 22, 2020, for India and other top major affected nations and also predictions were made regarding the number of cases for India over the next 17 days i.e from 23 June 2020 to 9 July 2020. Linear Regression model, Support Vector Machine Regressor (SVM) model, Autoregressive Integrated Moving Average (ARIMA) model and Facebook's Prophet model were used for prediction based on the Kaggle downloaded dataset with data collected from January 22, 2020, to June 22, 2020. By 22 June 2020, the disease has spread across more than 200 countries, reporting 12,322 confirmed cases, 45,26,333 recovered cases and 4,72,171 COVID-19 deaths. Assessment of this epidemic allows the Government to take the appropriate steps to curb the threat of this global pandemic.