GDA Models Wizard Extended Options - Advanced Tab

Select the Advanced tab of the GDA Models Wizard Extended Options dialog to access the options described here.

Element Name Description
Prior probabilities of classifications Use the options in the Prior probabilities of classifications group box to select the type of prior classification probabilities for your analysis. The prior classification probabilities are used in the classification of cases (observations) based on the current set of predictor variables; see also the Discriminant Analysis module for additional details.
Estimated Select the Estimated option button to compute the prior classification probabilities for the classes (categories or groups) specified in the dependent variable from the data. For example, suppose a dependent variable contains codes that specify two groups, and the data file contains 7 observations in group 1 and 3 in group 2 (and there are 10 observations in total), then this option will set the prior probabilities as equal 0.7 and 0.3, respectively.
Equal Select the Equal option button to assign equal prior classification probability to each group (class) in the dependent variable. For example, if the dependent variable specifies 3 groups, then the prior classification probabilities are set to 1/3 for each group.
User spec Select the User spec. option button to set your own prior probabilities for classes of the dependent variable. Click the button to display the Select Apriori Classification Probabilities dialog, which is used to specify the user defined probabilities. Note that this option will not be available until you have specified the codes for your dependent variable on the Quick tab.
Cross-validation Click the Cross-validation button to display the Cross-Validation dialog, which is used to specify a categorical variable and a (code) value to identify observations that should be included in the computations for fitting the model (the analysis sample); all other observations with valid data for all predictor variables and the dependent variable will automatically be classified as belonging to the validation sample; note that all observations with valid data for all predictor variables but missing data for the dependent variable will automatically be classified as belonging to the prediction sample (see the Cases tab and the Residuals tab for a description of available statistics for the prediction sample).
Model building options Select from the Model building options the way in which you want your discriminant analysis model to be built.
All effects Select the All effects option button to enter all effects specified in the current design (see Effects in design) into the regression equation.
Forward stepwise, Backward stepwise, Forward entry, Backward removal Select these option buttons to perform stepwise selection of predictor variables and effects. Forward options will cause variables to be moved into the model, backward options will start with a model with all predictor variables and effects in the model, which are then removed. The Forward stepwise and Backward stepwise options will at each step cause STATISTICA to consider simultaneously the addition or removal of a variable or effect, based on the current specifications of the p or F to enter/remove. See Stepwise options for additional details. The Forward entry and Backward removal options will only allow for variables or effects to be entered or to be removed, respectively, depending on the chosen method (forward or backward).

For example, if Forward stepwise is selected, STATISTICA will at each step consider both a step "forward", i.e., entry of another variable or effect into the model (based on the p or F to enter), and a step "backward", i.e., removal of a previously entered variable or effect from the model (based on the p or F to remove). The reason the Forward stepwise method usually adds rather than removes variables or effects (i.e., the reason why it is a forward selection method) is because of the required setting of the p or F,enter/remove values, which have to be specified so that p to enter is smaller (F to enter is larger) than the p to remove (F to remove), thus guaranteeing that significant predictor variables or effects are entered into the model, and not removed. Most of the widely used algorithms for stepwise selection of predictor variables use the Forward stepwise and Backward stepwise methods.

Note: you can force selected effects into the model (i.e., they will always be entered, and never be removed); specifically, when the Effects to force value (see below) k is greater than 0 (zero), then the first k effects in the design (see Effects in Design) will be forced into all models that are evaluated.
Best subsets Select the Best subsets option button to perform a search of all possible subsets of effects specified in the current analysis design. When this option button is set, various additional options will become available for steering the search for the best subset; see Best Subset Options for details. As discussed in the context of best subset regression in General Regression Models (GRM), the total number of all possible subsets (that need to be reviewed by STATISTICA) can become excessively large, when there are many effects in the model, and many large subset sizes are being considered (via the Start and Stop options). Thus, carefully review the best subset options before clicking OK. Note that you can force selected effects into the model see Effects to force below for further details.
Effects to force The Effects to force field allows you to force selected effects into the model (i.e., they will be part of every model that is considered); specifically, when the Effects to force value k is greater than 0 (zero), then the first k effects in the design (see Effects in design) will be forced into all models that are evaluated. Refer to the GRM Introductory Overview for a discussion of best subset regression. This option is available only if either one of the stepwise model building methods (Forward, Backward, etc.) or Best subsets is selected.
Sweep delta Enter the negative exponent for a base-10 constant delta (delta = 10-sdelta) in the Sweep delta field; the default value is 7. Delta is used (1) in sweeping, to detect redundant columns in the design matrix, and (2) for evaluating the estimability of hypotheses; specifically a value of 2*delta is used for the estimability check.
Inverse delta Enter the negative exponent for a base-10 constant delta (delta = 10-idelta) in the Inverse delta field; the default value is 12. Delta for matrix inversion is used to check for matrix singularity in matrix inversion calculations.

For alternative ways of specifying designs in GDA, see also Methods for Specifying Designs.