PREDICTION OF DEATHS FROM COVID-19 IN NIGERIA USING VARIOUS MACHINE LEARNING ALGORITHMS

  • O. H. Onyijen Department of Mathematical and Physical Sciences, Samuel Adegboyega University, Ogwa.
  • A. Hamadani Sher-e-Kashmir University of Agricultural Sciences and Technology of Kashmir, India.
  • S. Awojide Department of Mathematical and Physical Sciences, Samuel Adegboyega University, Ogwa.
  • I. E. Ebhohimen 3Department of Chemical Sciences, Samuel Adegboyega University, Ogwa, Nigeria.

Abstract

Machine learning techniques are deployed to process large datasets and can be used for predicting epidemiological events. In this study, machine learning models were used to analyse datasets from the Nigeria Center for Disease Control (NCDC) to explain the effect of total cases and other variables on total deaths from February 29, 2020 to July 20, 2020. Simple Linear Regression Model correctly predicted total deaths from total cases with a R2 value of 1 and it showed that the variable considered were significant. It was able to predict that if the total covid-19 cases were raised to 40,000, the predicted total number of deaths will be 924. The Multiple Linear Regression showed a good prediction with an R2 value of 0.99 and a RMSE of 0.136. The performance of the neural network model was evaluated using MAE and MSE, which indicated 0.0264 and 0.0013 respectively. Although the predicted total death toll in Nigeria did not manifest, the findings of this study indicate that total death is dependent on total cases. This study reveals the critical importance of prevention of the spread of this disease by all measures possible. This may be achieved by deploying relevant non-pharmaceutical interventions.

Published
2021-12-16
How to Cite
ONYIJEN, O. H. et al. PREDICTION OF DEATHS FROM COVID-19 IN NIGERIA USING VARIOUS MACHINE LEARNING ALGORITHMS. SAU Science-Tech Journal, [S.l.], v. 6, n. 1, p. 109-117, dec. 2021. ISSN 2659-1529. Available at: <https://journals.sau.edu.ng/index.php/sjbas/article/view/491>. Date accessed: 28 june 2022.
Section
Articles