Limitations:
First, this study predicted the total number of cases in the USA, where
each country has its’ own number of cases. Prediction for individual
countries needs different ARIMA models for each country. Second,
Covid-19 is dynamic, using data from different time periods for model
building and validation may lead to models with different AR, MA, and
differencing components with different validations. Third, we don’t know
if we used the variables from the whole or just the tip of the iceberg
of cases, but we know that we used all confirmed cases which is more
scientifically sound.