GCHAID Results - Classification Tab

Select the Classification tab of the GCHAID Results dialog box to access the options described here. This tab is available only if the Categorical response check box is selected on the Quick tab of the General CHAID Models Quick Specs dialog box.

Element Name Description
Sample Select Analysis, Prediction, Test, or Surrogate under Sample to specify for which type of sample to compute the predicted and residual statistics (classifications). You can display and plot predicted and residual values for all observations that were used to compute the current results (select Analysis), all observations that were not used to compute the current results, but have valid data for all predictor and dependent variables (select Test), or all cases that have valid data for the predictor variables, but missing data for the dependent variable (select Prediction). If these options are not available, no valid cases were found in the respective type of samples.
Predicted vs. observed by classes. This button is only available for classification-type problems. Click this button to produce a spreadsheet and a 3-D histogram of the predicted by observed classification frequencies.
Misclassification cost matrix Click this button to display the spreadsheet containing the user-specified costs of misclassifying cases or objects in each observed class of the dependent variable (columns) as another class (rows; all cost values will be 1 by default, i.e., if not altered by the user). These values will be used to adjust the risk estimate for the tree, available on the Summary tab.
Options for gain Gain is a measure of importance of a node that helps identify desired terminal nodes. Gain is calculated as the number or percent of cases in the selected category or as the average profit for each terminal node. The options here can be used to determine how the gain should be computed for each node. Note that these options are available only when performing classification analyses (i.e., for the categorical dependent variables - the Categorical response check box is selected on the Quick specs dialog box - Quick tab).
Average profit Select this option button to compute the gain as the weighted average of the number of cases in each category for the terminal node. All weights are 1 by default, but can be specified via the Profit option (see below).
Profit Click this button to display the Enter values of profit for the categories of dialog box. In this dialog box you can specify profit values (weights) to be used when computing the gain for each node. This button is only active if the Average profit option has been selected. By default, the profit values for all categories are equal to 1.
Percent of cases Select this option button if you want the gain to be computed as percent of cases of the response variable, for a selected category.
Category Select the category on which to base the percent-of-cases computations from this list. This option becomes available when Percent of cases is selected.