STATISTICAL MODELLING OF COVID-19 CASES IN NIGERIA WITH A NEGATIVE BINOMIAL AUTOREGRESSIVE MODEL

Abstract

Coronavirus disease (COVID-19) is a deadly global pandemic caused by a virus of the family coronaviridae. It is an infectious disease which affects respiratory systems and causes people to experience mild to moderate symptoms and sometimes severe cases of the disease which usually resulted into death especially among those patients with other comorbidity conditions and elderly with immunosenescence effects. Nigeria registered its index case of COVID-19 on 27^th February 2020. Subsequently, the number of reported cases were on increasing trend. Numerous studies on modelling the sporadic increase cases or spread of SARS-COV-2 using different methodologies were documented in literature. However, issues relating to over-dispersed problem and the presence of autocorrelation were not well handled in such methods. In this present study, the modelling of the spread of SARS-CoV-2 in Nigeria was done using a Negative Binomial Autoregressive model. Study data were collected on a daily basis from the update released by the Nigeria Centre for Disease Control from 1^st April, 2020 to 29^th May, 2021. The results showed that the number of confirmed SARS-CoV-2 cases increased comparatively between April-2020 to June-2020. However, the number of reported cases dropped steadily between July-2020 to Nov-2020. The data were over-dispersed and the presence of autocorrelation was observed. The results revealed that among the four NBAR estimated candidate models, NBAR (1) returned the lowest Akaike Information Criterion. Thus, NBAR (1) is the most parsimonious NBAR model for the data. Therefore, NBAR (1) can be used in predicting daily cases of SARS-CoV-2 in Nigeria