Dynamic Forecasting of COVID-19 Spread Using ARMAX Models

Abstract

The COVID-19 pandemic, unprecedented in the last 100 years, struck with the massive impact it had on humanity in multiple ways, but from the medical perspective the unpredictability of the course of the disease and the wide spectrum of clinical manifestations, with extreme serious and lethal accents for a percentage of patients, was a massive reason for concern. Therefore, researchers are still trying to decipher the body’s immune response and the duration of protection against a new reinfection. Immunological memory is the basis of durable protective immunity after infections or vaccinations. Its duration after infection with coronavirus 2 (SARS-CoV-2) is still under study, but more and more data are coming to complete this puzzle. Accurately foreseeing the virus’s progress is critical for allocating medical resources, managing industry operations, and maintaining economic growth. To this goal, dependable and appropriate forecasting techniques based on strong mathematical models are required. This paper investigates the use of ARMAX (Autoregressive Moving Average with Exogenous Inputs) models to forecast epidemiological trends in COVID-19 outbreaks, with an emphasis on Romania.