A Survey of Recent Studies on COVID-19 Outbreak Prediction Using Statistical and Machine Learning Methods

Çetin U. ,., Abut F.

Niğde Ömer Halisdemir Üniversitesi Mühendislik Bilimleri Dergisi, cilt.11, sa.3, ss.484-495, 2022 (Hakemli Dergi) identifier

Özet

COVID-19 is an infectious disease first discovered in Wuhan City, China, in December 2019. Ever since, COVID-19 has infected more than 70 million people and caused more than 1 million deaths worldwide. There is a need for models that predict the COVID-19 outbreak as accurately as possible to combat such an infectious and deadly disease. By using the results of the prediction models, governments can make better decisions and control measures about the disease, such as arranging budget and facility planning to combat the disease, deciding on how many medicines and medical equipment should be produced or imported, and how much medical staff is going to be needed. Consequently, various regression and classification models have been proposed for time series or supervised prediction of the COVID-19 outbreak in several countries and continents. This study aims to give an overview of recent studies on predicting the COVID-19 outbreak utilizing statistical and machine learning methods. Particularly, for each study, we outline the utilized groundtruth dataset characteristics, the type of the developed models, the predictor variables, the statistical and machine learning methods, the performance metrics, and finally, the major conclusion. The survey results reveal that machine learning methods are promising tools for making predictions for various needs, such as predicting whether a patient is infected with COVID-19 or not, predicting the trend of COVID-19 outbreaks, or predicting which age groups are most affected by COVID-19.