TIME SERIES MODELING AND FORECASTING OF THE COVID-19 PANDEMIC in TURKIYE USING MULTI GENE GENETIC PROGRAMMING


Creative Commons License

Arslan S. , Öztürk C.

III. International Conference on COVID-19, Ankara, Turkey, 25 - 27 December 2020, pp.1-8

  • Publication Type: Conference Paper / Full Text
  • City: Ankara
  • Country: Turkey
  • Page Numbers: pp.1-8

Abstract

The widespread outbreak of Coronavirus disease 2019 (COVID- 19) has impacted almost all people in the world. Analysis and modelling of time series of COVID-19 are needed to refine the measures and control the pandemic and improve disease forecast. A lot of research has been done to forecast the global pandemic since the world faced. Turkiye, which has with the highest percentage of foreign visitors is one of the world's top ten destination countries, so the impact and future behavior of COVID-19 in the country is really important. The main objective of this paper is to extract forecasting models of the spread of coronavirus in Turkiye in terms of confirmed cases. To the best of our knowledge, this is the first attempt in the literature that aims to forecast the number of confirmed cases using automatic programming method in Turkiye. Multi Gene GP is a high-level automatic programming method that increases the efficiency of the standard GP with a multi-tree modeling structure. The data of the previous 14 days were used to forecast the 15th day with the models which are extracted by Multi Gene Genetic Programming (Multi Gene GP). Since the pandemic reached in Turkiye on 11 March, all pandemic-related time series were trained dates between 11 March – 19 December and we forecasted the next 14 days. Experimental results show that, the proposed Multi Gene GP extracted successful forecsting models and had the ability to forecast time series of COVID-19 in Turkiye.