All Times ET
Keywords: COVID-19, INGARCH, Negative Binomial, Parametric Bootstrap, Conditional Distribution
In recent times a lot of studies have been conducted on modelling COVID-19 in the confine of Mathematical Sciences. The area considered included prediction and detection using various techniques such as Non-linear differential equation, Bayesian techniques, Machine learning, to mention but a few. This study considers modelling cases and predicting new cases of COVID-19 using integer-valued generalised autoregressive conditional heteroscedasticity (INGARCH) process to predict the future daily cases of COVID-19 using daily data of Nigeria from 30th March 2020 to 20th September 2020. Poisson and Negative Binomial distribution based on INGARCH were used, and Negative Binomial distribution was found to outperform the Poisson model. Therefore, the result for Negative Binomial distribution was considered conditional distribution and parametric bootstrap method; and as such was used to forecast for future new cases which could be used to determine period to flatten the curve of coronavirus in Nigeria