Final Project:
Find two time series data sets of interest to you. You might use the Time Series Data Library
http://www-personal.buseco.monash.edu.au/~hyndman/TSDL/
to find your data. For each data set determine an appropriate
ARIMA model.
- Plot
the time series with the appropriately labeled axes. In words describe the Trend,
Seasonality, and Irregular components of the time series. Plot the ACF and PACF of the time
series.
- Is the
time series stationary? Does it
have constant variance? Constant
variance?
- If the
time series does not have constant variance, suggest a transformation and
plot the transformed time series.
- Plot
the ACF and PACF of the transformed time series. Examine these plots for
nonstationarity. If appropriate
difference the data and seasonally difference the data. (Note this determines d and D in the ARIMA(p,d,q)(P,D,Q)s model.)
- After
differencing the data plot it and the ACF and PACF. Examine these plots for the seasonality
parameters P and/or Q. If
appropriate apply the model and save the residuals.
- Plot
the residuals and the ACF and PACF of the residuals. Examine these plots for nonseasonal
parameters p and/or q. If
appropriated apply the model and save the residuals.
- Plot
the residuals and the ACF and PACF of the residuals. Do the residuals appear to be white
noise?
- Report
the formula with the estimated parameters for your model. Test each parameter for statistical
significance.
- Forecast
your model one season into the future using you fitted ARIMA model.