ANALYSIS OF COVID-19 DATA AND PREDICTING FUTURE CORONAVIRUS CASES BY USING MACHINE LEARNING ALGORITHMS

Abstract

Background: The Covid-19 pandemic has posed significant challenges to healthcare systems worldwide. Effective strategies to manage the pandemic require accurate and timely forecasting of the spread of the virus. Machine learning (ML) algorithms offer a promising approach for predicting the number of Covid-19 cases. Objectives: This thesis work aims to analyze the Coronavirus data, and the number of cases and predict the future behavior of Covid-19 in Kazakhstan which helps to make key decisions related to the virus and prevent the country from the global economic crisis. Methods: The study utilized publicly available data sources to create a comprehensive Covid-19 dataset. The dataset included daily counts of confirmed Covid-19 cases, deaths, recoveries, and tests across multiple countries and regions worldwide. This work used four ML algorithms in our study, including a decision tree, random forest, linear regression (LR), and polynomial regression. Evaluation of the performance of the models based on r2 score, MAE, MSE. Results: Results showed that all four ML algorithms produced reasonably accurate predictions of Covid-19 cases. The random forest and decision tree algorithms outperformed the other models, with an accuracy rate of over 85% and 90% respectively. The linear and polynomial regression models had accuracy rates of approximately over 75%. Conclusion: In conclusion, this study demonstrates the potential of ML algorithms for predicting the number of Covid-19 cases. Findings suggest that the random forest algorithm is the most effective in forecasting Covid-19 cases. The results of this study may help inform policymakers and healthcare professionals in developing effective strategies to manage the Covid-19 pandemic.