Analyzing and Predicting Covid-19 epidemic using Machine Learning Techniques

Abstract

The entire world is suffering from a novel disease called covid-19 caused by a coronavirus since 2019. The main reason for the seriousness of the disease is the lack of efficient legitimate medication or vaccine. The World Health Organization (WHO) suggested several precautions to regulate the spread of disease and to reduce the contamination thereby reducing deaths. In this paper, we analysed the covid-19 dataset available in Kaggle. The previous contributions from several authors of similar work focused on covid-19 datasets having a limited number of samples. Our paper used the dataset updated till November 15th 2020. Three different aspects are considered mainly in this paper, namely the number of confirmed cases, number of recovered cases, and number of death cases. All the aspects are analysed in a daily and weekly manner. We applied linear regression, polynomial regression, and holt’s method to predict the future number of confirmed, recovered, and death cases. This analysis is useful for the health sectors and frontline workers to help reduce the contamination caused by this disease.