Publication date: 08/13/2020

Stepwise Platform Options

The Stepwise Fit red triangle menu contains the following options:

K-Fold Crossvalidation

(Available only for continuous responses.) Performs K-Fold cross validation in the selection process. When selected, this option enables the Max K-Fold RSquare stopping rule (Stepwise Regression Control Panel). For more information about validation, see Using Validation.

All Possible Models

(Available only for continuous responses.) Fits all possible models up to specified limits and shows the best models for each number of terms. Enter values for the maximum number of terms to fit in any one model. Also enter values for the maximum number of best model results to show for each number of terms in the model. Categorical variables are represented using indicator variables. See Models with Nominal and Ordinal Effects. You can restrict the models that appear to those that satisfy strong effect heredity. See The All Possible Models Option.

Model Averaging

(Available only for continuous responses.) Enables you to average the fits for a number of models, instead of selecting a single best model. See The Model Averaging Option.

Plot Criterion History

Creates a plot of AICc and BIC versus the number of parameters. The Criterion History plot contains two shaded zones. Define the minimum AICc value as Vbest. The green zone is defined by the range [Vbest, Vbest+4]. The yellow zone is defined by the range (Vbest+4, Vbest+10].

Plot RSquare History

(Available only for continuous responses.) Creates a plot of training and validation R-square versus the number of parameters.

Clear History

Clears and resets the step history.

Export Model with Validation

(Available only when you have entered a Validation column in the Stepwise launch window.) Adds the Validation column to the Model Specification window when you select Make Model. Runs the model with the Validation column when you select Run Model.

This option is selected by default.

Model Dialog

Shows the completed Fit Model launch window for the current model.

Want more information? Have questions? Get answers in the JMP User Community (community.jmp.com).
.