System for Forecasting COVID-19 Cases Using Time-Series and Neural Networks Models

Abstract

COVID-19 is one of the biggest challenges that countries face at the present time, as infections and deaths change daily and because this pandemic has a dynamic spread. Our paper considers two tasks. The first one is to develop a system for modeling COVID-19 based on time-series models due to their accuracy in forecasting COVID-19 cases. We developed an “Epidemic. TA” system using R programming for modeling and forecasting COVID-19 cases. This system contains linear (ARIMA and Holt’s model) and non-linear (BATS, TBATS, and SIR) time-series models and neural network auto-regressive models (NNAR), which allows us to obtain the most accurate forecasts of infections, deaths, and vaccination cases. The second task is the implementation of our system to forecast the risk of the third wave of infections in the Russian Federation.

Time series analysis on Alpha, Delta, and Omicron waves of COVID-19 infections in Indian provinces

“…This demonstrates that no lagged values of the variable under study are required, as the two lagged values of the error have been used and the series has been differenced twice to make it stationary for better forecasting. Similar to our findings, Duan and Zhang (2020) found the best-fitted ARIMA models to be ARIMA (6,1,7) in Japan and ARIMA (2,1,3) in South Korea for confirmed COVID-19 cases (11), while Singh et al (2020) found ARIMA (0, 1, 0) to be the best fitted ARIMA model for forecasting daily confirmed COVID-19 cases in Malaysia (25). Kufel (2020) discovered that ARIMA (1, 2, 0) was the best-fitting ARIMA model for forecasting the dynamics of confirmed COVID-19 cases in a subset of European countries (17).…”

“…Furthermore, the series has been differencing twice or integrated for the second order to make it stationary. Similarly, the best-fitted ARIMA based on ACF and PACF for the recovered cases is the ARIMA (5,2,4), which shows that 5 lagged values of the variable under consideration, 4 lagged values of the errors, and integration for the second-order have been used to make it stationary for the best fitting and forecasting. Figure -4(c) of the ARIMA model for the deceased cases shows that ARIMA is the best-fitted ARIMA model (0, 2, 2).…”

An Application of Neural Networks to Predict COVID-19 Cases in Italy

“…Khan and Gupta [3] chose an ARIMA (1,1,0) model for predicting Indian COVID-19 infection cases considering data that followed a linear trend. Abotaleb and Makarovskikh [4] proposed a combined ARIMA, Exponential Smoothing, BATS, and TBATS hybrid model for data collected until March 2021 in Russia. Gecili et al [5] proposed an ARIMA model for American and Italian data collected from February 2020 until April 2020.…”

Time Series Analysis of COVID-19 Data- A study from Northern India

“…However, no universal strategy for selecting models for forecasting COVID-19 spread has been established. (8)…”