Abstract
India now ranks 3rd in the list of countries affected due to the COVID-19 pandemic. With more than 0.7 million confirmed COVID–19 cases and a gradual withdrawal of nation-wide lockdown, it is expected India may see a further surge in the number of cases. So, predicting the expected number of cases is the need of the hour. Most of the existing methodologies work well for the short term but perform poorly when it comes to predicting the long term. Hence, in this study, we propose a novel strategy for the prediction of COVID–19 cases by employing Multiple Aggregation Prediction Algorithm (MAPA) with two different ways of predictions called the Principal and the Exponential Predictions. Exponential Prediction has been performed using MAPA on the number of cases, derived from predicted R0. From the Principal and Exponential prediction, a third prediction called Mean prediction was derived by averaging both. We have validated this strategy using data of the different states of India and show that Principal, Mean, and the Exponential Predictions together capture a range in which the actual number of cases lies. We have used this strategy to forecast the range of COVID-19 cases for 45 days.