Predicting COVID-19 spread using machine learning algorithms (CROSBI ID 702216)
Prilog sa skupa u zborniku | izvorni znanstveni rad | međunarodna recenzija
Podaci o odgovornosti
Žmuk, Berislav ; Jošić, Hrvoje
engleski
Predicting COVID-19 spread using machine learning algorithms
COVID-19 was declared as a world health emergency in January 2020. Since then it has affected all aspects of our lives. Countries closed their borders, put their population in self-quarantine and closed businesses and schools. As of March 31, the infection sickened more than 770, 000 people all around the world with thousands of fatalities. There is a little consensus how long will the infection last and what the number of infeced will be. Therefore, it is essential to implement suitable methods for COVID-19 spread prediction which is the goal of this paper. An open source machine learning software Weka and its algorithms (Linear regression, Gaussian Processes, SMOreg and neural network Multilayer Perceptron) have been used to predict the number of cases and fatalities of COVID-19 disease for 10 days in the future. The accuracy of the forecasts is measured using MAPE and RMSE error metrics. The results of the analysis have shown that Weka and its algorithms can be successfully used for prediction of COVID-19 spread in the world. The results of the analysis indicated that the Gaussian processes and Multilayer perceptron neural network are the most precise algorithms for the prediction of new and total cases of COVID-19 disease on a global scale and on an individual country level. The values of MAPE criterion for 12 selected countries, in majority of cases, have shown a highly accurate or good forecasting ability. The results obtained from this analysis can be important for global community and especially for economic and health policy makers in order to guide COVID-19 surveillance and implement public health policy measures.
COVID-19 ; machine learning ; disease spread prediction
Prilikom objave rada u zborniku radova došlo je do greške u tehničkom oblikovanju rada te su izbrisane oznake na osima na svim grafičkim prikazima u radu. Stoga je priložena i ispravljena inačica rada. ISBN (hard copy) 978-953-7813-62-8 ISBN (on line-version) 978-953-7813-63-5
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
Podaci o prilogu
233-246.
2021.
objavljeno
Podaci o matičnoj publikaciji
Contemporary Economic and Business Issues
Drezgić, Saša ; Host, Alen ; Tomljanović, Marko ; Žiković, Saša
Rijeka: Ekonomski fakultet Sveučilišta u Rijeci
978-953-7813-62-8
Podaci o skupu
Nepoznat skup
predavanje
29.02.1904-29.02.2096