Download Full Text (289 KB)
With the onset of the COVID-19 pandemic, it has become of increasing interest to both monitor and predict the growth of its infection rates. In order to analyze the accuracy of epidemiological prediction, we consider two different models for prediction, the Susceptible Infected and Removed (SIR), and Autoregressive Integrated Moving Average (ARIMA) models. Using a dataset of Clark County COVID-19 infections, we create various ARIMA and SIR models that attempt to predict the progression of COVID-19 infections whilst comparing these predictions to the dataset. We observed that the ARIMA model performed more accurately overall, having a much lower Root Mean Squared Error than its counterpart.
Infection; Epidemiology; Machine learning; COVID-19
Applied Mathematics | Epidemiology | Numerical Analysis and Computation
Domantay, Janelle; Pivavaruk, Ilya; and Taksheyev, Victor, "Modeling COVID-19 Infection Rates using SIR and ARIMA Models" (2021). Undergraduate Research Symposium Posters. 6.