Title: Modelling of COVID-19 spread time and mortality rate using machine learning techniques

Authors: Ahmad Arrabi; Amjed Al-Mousa

Addresses: Computer Engineering Department, Princess Sumaya University for Technology, Amman, Jordan ' Computer Engineering Department, Princess Sumaya University for Technology, Amman, Jordan

Abstract: One of the main issues in dealing with the COVID-19 global pandemic is that governments cannot predict the time it spreads or the mortality rate. If known, these two factors would have helped governments take appropriate measures without being excessively cautious and negatively impacting populations' mental health and economic outcomes. This paper presents a machine learning (ML)-based model that helps assess the rate at which the virus spreads in a country as well as the mortality based on multiple health, social, economic, and political factors. The method predicts how long a country's cases take to reach 5%, 10%, 15%, and 20% of its population. The prediction was conducted by regularised linear regression models and support vector machine regression (SVR). The SVR model achieved the highest median accuracy of 97%. Meanwhile, the ridge regression model achieved the best median accuracy of 84% for predicting the mortality rate.

Keywords: COVID-19; machine learning; support vector machine regression; SVR; kernel; regularised linear regression; leave one out cross-validation; LOOCV.

DOI: 10.1504/IJIIDS.2023.131412

International Journal of Intelligent Information and Database Systems, 2023 Vol.16 No.2, pp.143 - 166

Received: 12 Sep 2022
Accepted: 11 Mar 2023
Published online: 09 Jun 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article