·
Select an economic time
series that you find interesting. Say something about why you find it
interesting to analyze the series.
·
While the Federal Reserve
Economic Data (FRED) site at the St. Louis Fed is a useful source, you are
welcome to use other sources also.
·
Clearly state the time
period (e.g., July 1995 through Dec 2021) that is being covered.
·
If there are specific
sub-periods when the series was expected to behave unusually (e.g. , the
pandemic or the War in Ukraine) you may want to use appropriate dummy variables
for it.
·
Save some of your later
period observations for model validation through forecasting while using the
earlier sub-period for estimation.
·
Use your model to
forecast the series for the validation period.
·
Explain how you decided
to select the ARIMA structure of your model – e.g., AR or MA, or Mixed ARMA;
how many lags, etc.
·
Compare your ARIMA model
with a time series regression model. For example, you may try to do a
regression of miles driven on the price of gasoline as a time series
regression. Alternatively you may want to do an univariate ARIMA Model of miles
drive.
·
Summarize your results
and your main conclusions.