Modeling The Fatality Of Covid-19 Pandemic In Nigeria Using Some Regression Models
Edike Nnamdi
Department of Mathematics and Statistics, Ambrose Alli University, P.M.B. 14, Ekpoma, Edo State, Nigeria.
Braimah Joseph Odunayo
Department of Mathematics and Statistics, Ambrose Alli University, P.M.B. 14, Ekpoma, Edo State, Nigeria.
Umar Shehu Salisu
Department of Statistics, Auchi Polytechnic, Auchi, Edo State, Nigeria
Ogbeide Efosa Michael
Department of Mathematics and Statistics, Ambrose Alli University, P.M.B. 14, Ekpoma, Edo State, Nigeria.
Keywords: COVID-19 pandemic, model, linear regression, log-log regression, coefficient of determination
Abstract
The outbreak of corona virus (covid-19) pandemic stirs up a lot of responses from researchers all over the world. Statisticians have also deployed useful statistical tools and techniques to provide useful information on the spread of the pandemic to enable state actors respond accordingly. This paper models the fatality of covid-19 pandemic in Nigeria. The data was retrieved from the database of Our World in Data (OWID). A scatter plot of the data revealed an approximately linear relationship between the cumulative total confirmed cases and the cumulative total covid-19 related deaths in Nigeria. Linear regression models involving the logarithmic transformation of the variables were employed in this study. The models include the simple linear, the linear-log, the log-linear, and the log-log regression models. The performances of the various models were compared using the coefficient of determination (R2). The result showed that, though the studied models were significant at a 0.05 level, the log-log regression model with a coefficient of determination of 0.9780 provides the best fit and hence a better prediction power to the data.