* PS10.5, using DATA10-2 for the application in Section 10.6 * SAMPLE 1 744 READ(data10-2) DAYHOUR LOAD TEMP GENR LOAD1=LAG(LOAD) * * Estimate the model with a lagged dependent variable. * SAMPLE 2 744 OLS LOAD LOAD1 TEMP / RESID=UT * * Auxiliry regression for testing AR(24). * SAMPLE 2 744 OLS UT LOAD1 TEMP UT(1.24) GEN1 LM=$N*$R2 DISTRIB LM / TYPE=CHI DF=24 * * The results below do not match those found in the textbook. ESL uses the * two-step iteration discussed in Section 9.5 of the book. SHAZAM uses a * Gauss-Newton non-linear algorithm. The qualitative results, however, are * the same. * SAMPLE 2 744 AUTO LOAD LOAD1 TEMP / ORDER=24 * * The NUMARMA= option specifies the number of AR or MA coefficients to be * estimated if some of the coefficients are to be restricted to zero. * AUTO LOAD LOAD1 TEMP / ORDER=24 NUMARMA=18 1 2 5 6 7 8 9 12 13 14 15 17 18 19 21 22 23 24 AUTO LOAD LOAD1 TEMP / ORDER=24 NUMARMA=17 1 2 5 6 7 9 12 13 14 15 17 18 19 21 22 23 24 AUTO LOAD LOAD1 TEMP / ORDER=24 NUMARMA=16 1 2 5 7 9 12 13 14 15 17 18 19 21 22 23 24 AUTO LOAD LOAD1 TEMP / ORDER=24 NUMARMA=15 1 2 5 7 9 12 13 14 15 17 19 21 22 23 24 * DELETE / ALL STOP