Computing for Chapter 2 of Box, Jenkins, & Reinsel.

  To solve the homework exercises for this chapter, you should be able to

(1)

Produce a time series plot, i.e. plot an observed series z_t, $t=1, \ldots, n$, as a function of time t.

(2)

Produce a lagged scatterplot, i.e. plot z_t+h vs. z_t for some lag h, typically h = 1 or 2, but not much more.

(3)

Calculate and plot the sample autocovariance and autocorrelation, and plot the requisite bands to indicate significant sample autocorrelations.

(4)

Compute and plot a sample spectrum (periodogram).

We will defer the spectrum to another day and only treat the first 3.

We carry this exercise out with the cow temperature data set. This is in the file ``cow.dat'' which you may download from the same place I got it:

http://www-personal.buseco.monash.edu.au/ hyndman/forecasting/welcome.htm

go to the "Data" link on the sidebar, and the cow data is under Exercise 2.4.

In section 2, we show how to do this with SAS. In section 3, we show how to do this with Splus.