Statistical Analysis of COVID-19 Death Cases in Nigeria Using Machine Learning Approaches and Count Data Regression Models

Abstract

In this paper, different methods suitable for analyzing count data based on relevant data characteristics were studied using two of the most popular count regression models (i.e., Poisson regression and negative binomial regression) and some machine learning algorithms, namely decision tree, support vector machine, and neural networks. This study model confirmed cases of COVID-19 death in Nigeria. The data used in the study are the number of cumulative COVID-19 confirmed cases, the number of discharged patients, the number of active cases on admission as predictors, and the number of COVID-19 deaths as the response variable obtained from the Nigeria Centre for Disease Control (NCDC). It is observed that out of the three machine language algorithms considered, the support vector machine learning algorithm provides a better result and outperforms the count regression models in terms of minimum MSE, RMSE and MAE. Then, the support vector machine learning algorithm is recommended for modeling the dynamics of COVID-19 death in Nigeria.