Comparative analysis and forecasting of COVID-19 cases in various European countries with ARIMA, NARNN and LSTM approaches

İsmail Kırbaş; Adnan Sözen; Azim Doğuş Tuncer; Fikret Şinasi Kazancıoğlu

doi:10.1016/j.chaos.2020.110015

Comparative analysis and forecasting of COVID-19 cases in various European countries with ARIMA, NARNN and LSTM approaches

Chaos Solitons Fractals. 2020 Sep:138:110015. doi: 10.1016/j.chaos.2020.110015. Epub 2020 Jun 13.

Authors

İsmail Kırbaş¹, Adnan Sözen², Azim Doğuş Tuncer^{3

4}, Fikret Şinasi Kazancıoğlu⁵

Affiliations

¹ Department of Computer Engineering, Faculty of Engineering-Architecture, Burdur Mehmet Akif Ersoy University, Burdur, Turkey.
² Department of Energy Systems Engineering, Faculty of Technology, Gazi University, Ankara, Turkey.
³ Department of Energy Systems Engineering, Faculty of Engineering-Architecture, Burdur Mehmet Akif Ersoy University, Burdur, Turkey.
⁴ Institute of Natural and Applied Sciences, Gazi University, Ankara, Turkey.
⁵ Turkish State Railways, Ankara, Turkey.

Abstract

In this study, confirmed COVID-19 cases of Denmark, Belgium, Germany, France, United Kingdom, Finland, Switzerland and Turkey were modeled with Auto-Regressive Integrated Moving Average (ARIMA), Nonlinear Autoregression Neural Network (NARNN) and Long-Short Term Memory (LSTM) approaches. Six model performance metric were used to select the most accurate model (MSE, PSNR, RMSE, NRMSE, MAPE and SMAPE). According to the results of the first step of the study, LSTM was found the most accurate model. In the second stage of the study, LSTM model was provided to make predictions in a 14-day perspective that is yet to be known. Results of the second step of the study shows that the total cumulative case increase rate is expected to decrease slightly in many countries.

Keywords: ARIMA; COVID-19; Forecasting; LSTM; Modeling; NARNN.