Prediction of Spread of Covid-19 Infections - A Time-Series Based Approach
Abstract
Covid-19 has hard-pressed world to handle one of the major health emergencies of modern times. In this scenario, prediction or forecasting the new infections and the rate of death is the need of the hour for effective preparedness in terms of medical facility, vaccinations and other requirements to eradicate the disease from further spread. As on 29th August 2021, according to the Ministry of Health and Family Welfare, Govt of India Sources, a total of 4,37,830 Death cases have been reported in 28 states and 8 union territories. The vaccination to all, proper medical treatment policy, maintaining social distancing, cleanliness and awareness of seriousness about the infection will lead to the end of the current pandemic situation. This paper aims to describe the empirical study of modelling and forecasting time series data of COVID 19 for India. COVID data for the period of 30.01.2020 to 27.08.2021 were collected from the portal of Ministry of Health, Govt of India, analysed by applying Time Series approach and the ARIMA model is used for forecasting. The proposed model is tested using Lag 1 autocorrelation of error (acf1), minmax error and correlation and the obtained results are promising. The proposed time series models proved to be an effective approach as the level of accuracy is close to 96% in case of both infections and deceased rate.
