PERFORMANCE OF BOX-JENKINS ESTIMATION METHOD IN COVID-19 PREDICTIONS
PERFORMANCE OF BOX-JENKINS ESTIMATION METHOD IN COVID-19 PREDICTIONS

Summary/Abstract: The course of the COVID-19 pandemy, which affects the world, is one of the main focal points of our world. The predictions made regarding the future of this epidemic, which has made a great impact on the structure of our daily life, on the international system and deeply affected living conditions, are of great importance. These estimation methods which statistically carried out are of great importance in terms of clinical studies on COVID-19 and the policy of states that take measures to mitigate the course of the epidemic. When we look at these statistical estimates in publications and media, we can see that there are significant deviations between the numerical values of the estimates and the actual values. Due to factors such as the epidemiological nature of the virus, the difficulty of predicting the course of such cases, and the limited stochastic process, the performance of the estimates is generally low. In the study, in order to reveal the reason for the inconsistency in the COVID-19 estimates made in the epidemiological perspective, Box-Jenkins method used works done in the literature on the subject were examined. The deviations in the studies were analyzed and the reasons for the insufficiency of statistical methods in explaining the course of the virus in general were revealed. However, Box-Jenkins method of executing 10-day forecast on the course of Covid-19 in Turkey about the state of the received deviation tendency in the literature were investigated. According to the results of both the studies in the literature and the empirical analysis conducted within the scope of this study, significant deviations were observed between the predicted values made by the Box-Jenkins method and the actual values.

• Details
• Contents