You are given an annual time series with 40 consecutive values and asked to fit a fifth-order auto-regressive model.
a. How many comparisons are lost in the development of the auto-regressive model?
b. How many parameters do you need to estimate?
c. Which of the original 40 values do you need for forecasting?
d. State the fifth-order auto-regressive model.
e. Write an equation to indicate how you would forecast x years into the future.
a. The dependent variable of a pth-order auto-regressive model is dependent on the p previous variables in the time series. So the comparison is only possible starting with (p+1)th term. …
The solution gives detailed steps on answering several questions on statistical forecasting. All formula and calculations are shown and explained.