Author(s): B Odoi, RA Ofosu and K William

Abstract: The World Health Organization (WHO) proclaimed COVID-19 a worldwide pandemic in December 2019, and a number of mathematical models have been developed to forecast its spread. Since March 14, 2020, when it reported its first incidence, COVID-19 has spread throughout Ghana as a whole. Despite the preventative and control measures that the government and stakeholders have subjected to and accepted, the illness continues to spread at an alarming rate. The goal of this work was to develop a generalized Poisson regression model to forecast the growth rate of COVID in Ghana over the next years as well as the country's outcome in terms of this pandemic. To conduct this study, a daily historical dataset of COVID-19 cases in Ghana from March 14, 2020, to August 26, 2022, was employed. The generalized Poisson linear regression model was the methodology employed in the investigation. Since it was assumed that the data would have a probability distribution, the log-linear model provided a good hypothesis for the data, and the estimators we had under the probability model's hypothesis were valid. The estimators and actual data were compared, and the log-linear model was evaluated. The analysis's findings showed that there was over-dispersion in the dataset since the variance of the recorded instances was higher than the mean. The residual deviation was bigger than the degree of freedom, according to a confirmatory study based on over-dispersion that was carried out once more. The over-dispersion outcomes from the generalized Poisson model were then corrected using the Quasi Poisson to make the model appropriate. The Quasi Poisson model showed that the number of COVID-19 cases was falling by 56.02% on a daily basis. Despite the fact that the mortality rate was rising by 98.7%. This demonstrates that the number of documented death cases each day is not a factor in death cases.

DOI: 10.22271/maths.2024.v9.i2b.1693

Pages: 131-136 | Views: 121 | Downloads: 31

Download Full Article: Click Here