Research Article Open Access

Robust Modelling of Coronavirus Cases in Egypt: Poisson INARCH and Negative Binomial INARCH

Hanan Elsaied1
  • 1 Suez Canal University, Egypt

Abstract

This study compares robust Poisson INARCH(P) models (more briefly: RP-INARCH) and robust negative binomial INARCH(p) models (more briefly: RNB-INARCH) to fit the new daily confirmed cases for the first wave of COVID 19 in Egypt. The robust estimation of these models is based on some modifications of the Conditional Maximum Likelihood Estimates (CMLE). The simulation results show that RNB-INARCH is more robust than RP-INARCH, but less efficient if the data contain isolated or patched additive outliers in terms of the bias calculation, whereby the low-bias model is more robust. These results are confirmed by the application study on COVID-19 data. The Akaike Information Criterion (AIC) is also compared for these models.

Journal of Mathematics and Statistics
Volume 17 No. 1, 2021, 50-58

DOI: https://doi.org/10.3844/jmssp.2021.50.58

Submitted On: 3 May 2021 Published On: 23 June 2021

How to Cite: Elsaied, H. (2021). Robust Modelling of Coronavirus Cases in Egypt: Poisson INARCH and Negative Binomial INARCH. Journal of Mathematics and Statistics, 17(1), 50-58. https://doi.org/10.3844/jmssp.2021.50.58

  • 142 Views
  • 78 Downloads
  • 0 Citations

Download

Keywords

  • COVID 19 Data
  • Robust Poisson INARCH(p) Models
  • Robust Negative Binomial INARCH(p) Models
  • Isolated Additive Outliers
  • Patch Additive Outliers