*
* Example from chapter 6, pp 265-291
*
open data bankdata.dat
calendar(m) 1988:2
data(format=free,org=columns) 1988:2 1993:1 eom aaa govt34
*
set deom = eom{-1}-eom
set d34  = govt34{-1}-govt34
*
linreg deom * 53
# constant aaa govt34 d34
*
graph(footer="Figure 6-7: Time Plot of Residuals")
# %resids
*
* The procedure REGCORRS graphs the autocorrelations of a series, generally used
* for residuals.
*
@regcorrs(footer="ACF of model residuals")
*
* The SEASONALS instruction creates a dummy seasonal variable It can create a
* dummy for any specific month, but the default is to create it as a dummy for
* the first month in the date scheme; in this case, February. (The data begin
* 1988:2). The other months are usually gotten by using leads of this. The first
* lead will be January, the second December, ... To get the set of dummies used
* in the regression in the text you want leads -1 to -11. (January through March).
*
seasonals seasons
linreg deom * 53
# constant aaa govt34 d34 seasons{-1 to -11}
graph(footer="Figure 6-8 Residuals from model with dummy variables")
# %resids
@regcorrs
*
set trend  = t
set trend2 = t**2
set trend3 = t**3
*
* The RATS instruction STWISE does stepwise regression (pp 285-287). The SLSTAY
* and SLENTER options fix a significance level to stay in the regression (in the
* backward pruning step) and to enter the regression (in the forward selection).
* These are set to the .10 values used in generating table 6-11 (the default
* values for both are .20). The FORCE=1 option forces the first regressor on the
* list (the CONSTANT) into all regressions.
*
stwise(method=stepwise,slstay=0.10,slenter=0.10,force=1) deom * 53
# constant aaa govt34 d34 seasons{-11 to -1} trend trend2 trend3