The current state of COVID-19 in Australia: importation and spread

Jessica Liebig; Raja Jurdak; Ahmad El Shoghri; Dean Paini

is the (which was not peer-reviewed) The copyright holder for this preprint 2 0 2 0 1 2 12 China Visitors 1 1 1 1 1 2 12 France Residents 0 1 1 0 1 2 12 France Visitors 1 0 1 1 1 1 12 Germany Residents 0 0 0 1 1 2 12 Germany Visitors 1 0 1 0 1 1 12 India Residents 0 0 3 0 1 1 12 India Visitors 0 0 3 0 1 1 12 Indonesia Residents 1 0 1 0 1 1 12 Indonesia Visitors 1 0 0 0 1 0 12 Iran Residents 0 1 1 0 1 1 12 Iran Visitors 1 0 2 0 1 1 12 Italy Residents 1 0 0 1 1 2 12 Italy Visitors 2 1 3 1 1 1 the SARIMA models used to predict passenger flows from different countries. Before finding the parameters for the SARIMA model, we transformed the individual time-series using a Box-Cox transformation to give the data a normal shape [17] . The transformed time-series was tested whether it is stationary using the Kwiatkowski-Phillips-Schmidt-Shin test [18] . To find the model with the best fit we perform a step-wise search over the model space [19] and chose the model with the lowest AIC (Akaike's Information Criterion). We considered the following parameter ranges: p ∈ [0, 5], d ∈ [0, 2], q ∈ [0, 5], P ∈ [0, 2], Q ∈ [0, 2]. All calculations have been carried out with Python's pmdarima statistical library.