Best Subset Regression

Click the Best subs button on the GLZ Analysis Syntax Editor - Keywords tab to display the Best Subset Regression dialog box, which contains keywords and options (specifications) for specifying a best subset analysis; see the description of GLZ Syntax for a list of all keywords (see also, GLZ Analysis Syntax Editor). Note that best subset regression is only available if the current model contains a single dependent variable. Click on a button to insert the respective keyword, option, symbol, or value at the current cursor location in the Analysis syntax edit field (in the GLZ Analysis Syntax Editor).

Keywords

The Keywords group box contains the following keywords:

BESTCRIT

BESTCRIT [=] { LIKELIHOOD };

  { SCORE };

  { AIC };

Example. BESTCRIT = SCORE;

Optional keyword used in conjunction with the MBUILD option bestsubset (note that this option is very similar to that described for the GRM syntax). Specify the criterion that is to be used for comparing models (with different subsets of effects) during the best subset computations. The default specification is SCORE (score statistic); see also the LIKELIHOOD and AIC (Akaike information criterion) options below.

Since the evaluation of the score statistic does not require iterative computations, best subset selection based on the score statistic is computationally faster, while the selection based on the other two statistics usually provides more accurate results. For more information see the Introductory Overview.

Applies to. GRM, GLZ (with different options)

MAXSUB

MAXSUB [=] Integer Value;

Example. MAXSUB = 10;

Optional keyword used in conjunction with the MBUILD option bestsubset. The integer value specified with this keyword will determine the number of subsets that will be displayed on the GLZ Results dialog, in the Model building spreadsheet; for example, if you specify 10 in this field, then you can later review (via option Model building button on the GLZ Results - Summary tab) the 10 best subsets according to the chosen criterion (see the keyword BESTCRIT above). The default value is 200.

Applies to. GRM, GLZ (implementation is slightly different; see the GRM syntax keyword MAXSUB)

Specifications

The Specifications group box contains specifications (options) that can be used in the analysis syntax. Note that the specifications discussed below are the ones that are unique to best subset regression. For details about the other specifications in this dialog box, see GLZ Analysis Syntax Editor - Specifications tab.

score

Click the score button to insert the specification score into the Analysis syntax edit field (in the GLZ Analysis Syntax Editor), at the current location of the cursor. This option is used in conjunction with the BESTCRIT keyword (see above, see also MBUILD) to specify that the score statistic is to be used when comparing the subsets of effects during best subset regression; note that this is the fastest (and hence default) BESTCRIT option.

likelihood

 Click the likelihood button to insert the specification likelihood into the Analysis syntax edit field, at the current location of the cursor. This option is used in conjunction with the BESTCRIT keyword (see above, see also MBUILD) to specify that the overall model likelihood is to be used when comparing the subsets of effects during best subset regression; note that the computation of the model likelihood for each subset can require some time; the fastest BESTCRIT option is SCORE.

AIC

Click the AIC button to insert the specification AIC into the Analysis syntax edit field, at the current location of the cursor. This option is used in conjunction with the BESTCRIT keyword (see above, see also MBUILD) to specify that the Akaike information criterion (AIC) is to be used when comparing the subsets of effects during best subset regression; note that the computation of the AIC statistic can require some time; the fastest BESTCRIT option is SCORE.

Close

Click this button to return to the Syntax Editor.

See also, GLZ - Index.