Selecting a Model

Option Name	Description
Model Selection
Selection method	By default, the complete model that you specified is used to fit the model. However, you can also use one of these selection methods: Forward selection specifies forward selection. This method starts with no effects in the model and adds effects. Backward elimination specifies backward elimination. This method starts with all effects in the model and deletes effects. Stepwise regression specifies stepwise regression, which is similar to the forward selection method except that effects already in the model do not necessarily stay there. LASSO specifies the LASSO method, which adds and deletes parameters based on a version of ordinary least squares where the sum of the absolute regression coefficients is constrained. If the model contains classification variables, these classification variables are split. Adaptive LASSO requests that adaptive weights be applied to each of the coefficients in the LASSO method. The ordinary least squares estimates of the parameters in the model are used in forming the adaptive weights.
Selection method (continued)	Elastic net specifies the elastic net method, which is an extension of LASSO. The elastic net method estimates parameters based on a version of ordinary least squares in which both the sum of the absolute regression coefficients and the sum of the squared regression coefficients are constrained. If the model contains classification variables, these classification variables are split. Least angle regression specifies least angle regression. This method starts with no effects in the model and adds effects. The parameter estimates at any step are “shrunk” when compared to the corresponding least squares estimates. If the model contains classification variables, these classification variables are split.
Add/remove effects with	specifies the criterion to use to determine whether an effect should be added or removed from the model.
Stop adding/removing effects with	specifies the criterion to use to determine whether effects should stop being added or removed from the model.
Select best model by	specifies the criterion to use to determine the best fitting model.
Selection Statistics
Model fit statistics	specifies which model fit statistics are displayed in the fit summary table and the fit statistics tables. If you select Default fit statistics, the default set of statistics that are displayed in these tables includes all the criteria used in model selection. Here are the additional fit statistics that you can include in the results: Adjusted R-square Akaike’s information criterion Akaike’s information criterion corrected for small-sample bias Average square error Bayesian information criterion Mallows’ Cp Press statistic, which specifies the predicted residual sum of squares statistic R-square Schwarz’s Bayesian information criterion
Selection Plots
Criteria plots	displays plots for these criteria: adjusted R-square, Akaike’s information criterion, Akaike’s information criterion corrected for small-sample bias, and the criterion used to select the best fitting model. You can choose to display these plots in a panel or individually.
Coefficient plots	displays these plots: a plot that shows the progression of the parameter values as the selection process proceeds a plot that shows the progression of the criterion used to select the best fitting model
Details
Selection process details	specifies how much information about the selection process to include in the results. You can display a summary, details for each step of the selection process, or all of the information about the selection process.
Add/remove classification effects	specifies which classification variables are included in the model as one or more actual variables. The number of variables is related to the number of levels of the classification variable. For example, if a classification variable has three levels (young, middle-aged, old), it might be represented by 3 variables. Each variable is a single degree of freedom effect. You can choose from these options: Add/remove as entire effect, which specifies that all or none of the variables for a classification effect are included in the model. Add/remove as individual single degree of freedom effects, which specifies that one or more of the individual variables are included in the model. Some individual variables might not be included in the model.
Model Effects Hierarchy
Model effects hierarchy	specifies how the model hierarchy requirement is applied and that only a single effect or multiple effects can enter or leave the model at one time. For example, suppose you specify the main effects A and B and the interaction AB in the model. In the first step of the selection process, either A or B can enter the model. In the second step, the other main effect can enter the model. The interaction effect can enter the model only when both main effects have already been entered. Also, before A or B can be removed from the model, the AB interaction must first be removed. Model hierarchy refers to the requirement that, for any term to be in the model, all effects contained in the term must be present in the model. For example, in order for the interaction AB to enter the model, the main effects A and B must be in the model. Likewise, neither effect A nor B can leave the model while the interaction AB is in the model.
Model effects subject to the hierarchy requirement	specifies whether to apply the model hierarchy requirement to the classification and continuous effects in the model or to only the classification effects.