Classification Trees Results - Cross-Validation Tab

Select the Cross-validation tab of the Classification Trees Results dialog box to access the options described here.

Learning sample. Use the options under Learning sample to control how cross-validation will be performed in the analysis.

Perform global CV. Click the Perform global CV button to display the Global Cross-Validation dialog box, in which you select options to perform global cross-validation.

v-fold for GCV. Use the v-fold for GCV box (and the accompanying microscrolls) to specify the number of cross-validation samples to be used in estimating the Global CV error.

Test sample. Use the options under Test sample to display either a Test Sample Misclassification Matrix or Predicted Classes spreadsheet. Note that these options are available only if a test sample has been specified for the analysis.

Misclassification matrix. Click the Misclassification matrix button to display the Test Sample Misclassification Matrix spreadsheet. The number of cases or objects in each observed class on the dependent variable (columns) misclassified as another class (rows) are displayed in the spreadsheet. The CV cost and the standard deviation are displayed in the spreadsheet header.

Predicted cases. Click the Predicted cases button to display the Predicted Classes spreadsheet for the test sample. The spreadsheet displays for each object in the test sample the object or case number (or case name, if used), the observed class, the predicted class, and the terminal node to which the object was assigned. Also shown for each object are the categories (for categorical predictors) or values (for ordered predictors) for each predictor variable.