Classification Trees Startup Panel - Advanced Tab

Select the Advanced tab of the Classification Trees Startup Panel to access options to select the variables that will be used in the analysis, the codes for the dependent variable and for any categorical predictor variables for classes or groups, and, where applicable, the sample identifier variable codes.

Variables. Click the Variables button to display the standard variable selection dialog box, in which the variables for the analysis are specified.

From the first list, select the dependent variable for the analysis. The dependent variable must contain text or numeric codes identifying the class, or group, to which each case or object belongs.

From the second list, select any categorical predictors for the analysis. Like the dependent variable, categorical predictor variables must contain text or numeric codes identifying the group to which each case or object belongs.

From the third list, select any ordered predictors (measured on at least an ordinal-level scale) for the analysis. Note that at least one categorical or ordered predictor variable must be specified for the analysis.

From the fourth list, an optional sample identifier variable can be selected. A sample identifier variable is used when the data file contains both a learning sample, from which the classification tree for the dependent variable is computed, and a test sample, which is used to test the predictive accuracy of the classification tree computed from the learning sample. The sample identifier variable must contain codes identifying the sample (learning or test) to which each case or object belongs.

Note: missing data. Statistica will apply casewise deletion of missing data; that is, cases will be deleted from the analysis if there are missing data on that case for any of the variables specified for the analysis. Thus, be careful when there are missing data present in the selected variables; the results for those variables without missing data may not be based on all available information (namely, cases where some other variable had missing data, causing the entire case to be dropped from the analysis).

Codes for variables. Use the options under Codes for variables to specify the codes for the dependent variable and for any categorical predictor variables for classes or groups to be included in the analysis.

Dep. variable. Click the Dep. variable button to display the Select Codes for Dependent Variable dialog box, from which the codes to be included in the analysis can be selected. The text or numeric codes for the variable specified as the dependent variable are used to construct the classes or groups to be analyzed.

Shortcut. If you leave all fields in the Select Codes for Dependent Variable dialog box empty (blank) and click the OK button, Statistica will identify and automatically use all available codes (for the previously specified dependent variable). Also, the same effect will be achieved if you do not explicitly specify any codes.

Cat. predictors. Click the Cat. predictors button to display the Select Codes for Categorical Predictors dialog box, from which the codes to be included in the analysis can be selected. If any categorical predictor variables have been specified for the analysis, the text or numeric codes for any such variables are used to construct the groups to be analyzed.

Shortcut. If you leave all fields in the Select Codes for Categorical Predictors dialog box empty (blank) and click the OK button, Statistica will identify and automatically use all available codes (for any previously specified categorical predictor variables). Also, the same effect will be achieved if you do not explicitly specify any codes.

Codes for samples. Use the boxes under Codes for samples to specify the sample identifier variable codes that uniquely identify the cases belonging to the learning and test samples if a sample identifier variable has been specified for the analysis.

Learning sample. If a sample identifier variable has been specified for the analysis, use the Learning sample box to specify the code for the sample identifier variable that uniquely identifies those cases belonging to the learning sample. By default, when a sample identifier variable has been selected, the Learning sample box will display the first valid code for the sample identifier variable that is found in the data file. If this is not the desired code for identifying the learning sample, double-click on the box (or press the F2 key on your keyboard) to display a Variable dialog box from which you can select the desired code from the list of valid variable values.

Test sample. If a sample identifier variable has been specified for the analysis, use the Test sample box to specify the code for the sample identifier variable that uniquely identifies those cases belonging to the test sample. By default, when a sample identifier variable has been selected, the Test sample box will display the second valid code for the sample identifier variable that is found in the data file. If this is not the desired code for identifying the test sample, double-click on the box (or press the F2 key on your keyboard) to display a Variable dialog box from which you can select the desired code from the list of valid variable values.