On the Temporal Analysis of COVID-19 Pandemic and Prediction of R 0

Background: The COVID-19 pandemic has affected millions of people and claimed numerous lives already. As of now, there are no available vaccines for the virus and it has become both imperative and challenging to forecast the COVID-19 cases, which will help to design effective clinical management and policy to fight the pandemic. Methods: With the objective to forecast the COVID-19 cases and Basic Reproductive Number (R0) country-wise, for more than a month ahead, we have adopted a data driven approach that employs Multiple Aggregation Prediction Algorithm (MAPA) for temporal predictions. Our strategy applies MAPA in two separate ways. First application is directly onto the number of cases and second is by calculation of R0 from the total number of cases, followed by application of MAPA. Findings: This novel workflow generates a Principal Prediction along with an Exponential Prediction that provides a range of values within which the total number of cases is expected to lie. The strategy and workflow have been validated for long term predictions (upto 45 days) with 51 countries showing Rising, Exponential growth and Plateauing number of cases, which contribute to at least 91% of the total number of cases in the world. Thereafter, we have made predictions of the possible number of COVID-19 cases that is likely to be witnessed in the next 45 days by these 51 countries, the world as a whole and the other 160 countries combined, that are affected by the pandemic. Interpretation: The integrative strategy developed in this study using machine learning and statistical tools can accurately predict the expected number of COVID-19 cases for the next 45 days with better accuracy as compared to other existing models and data driven approaches.Funding Statement: RRS acknowledges funding from Department of Science and Technology, Govt. of India (Sanction Order No. DST/ICPS/EDA/2018/General, dated 30.04.2019), PG acknowledges CSIR for Senior Research Fellowship. SN acknowledges DST INSPIRE Fellowship.Declaration of Interests: Authors declare no competing interests.