Using Machine Learning Techniques to Explore the Possibilities of Reducing the Spread of Corona Virus and its New Variants

Abstract

The Corona pandemic has been around for a while, and its threat to the world is growing. We believe that climate parameters and health prevention measures could be related to the number of reported Corona daily cases. In the literature there were different views on the nature of these relations using several datasets recorded from various parts of the world. In our research, data collected from zones with concentrated Corona cases: China, Europe and the United States were analyzed to understand the relation with climate as well as data at the global level to understand the relation with health prevention measures. Feature importance analysis revealed that temperature is the most important contributing attribute to the Corona cases' prediction models, followed by relative humidity. As well, the percentage of mask use and percentage of fully vaccinated individuals were found to have a great influence on the number of new Corona daily cases. The designed machine learning ensemble techniques had a maximum predication accuracy of 89.08%, and the produced possible interpretations for the designed models agreed with the performed feature importance analyses. We believe that the analysis approach followed in this research as well as the achieved findings could be very useful to other researchers who are interested in conducting more research investigation in the same research area on the new Corona variants. We also believe that policy makers could consider the findings of our research as they effectively plan their future health precautions measures to avoid further spread of the virus.