MISSING Subcommand (TREE command)
The MISSING subcommand controls the handling of nominal, user-missing, independent (predictor) variable values.
- Handling of ordinal and scale user-missing independent variable values varies between growing methods.
- By default, if the values of all independent variables are system- or user-missing, the case is excluded.
- Handling of nominal dependent variables is controlled by the DEPCATEGORIES subcommand.
- For the dependent variable, cases with system-missing or user-missing ordinal or scale values are always excluded.
NOMINALMISSING Keyword
The NOMINALMISSING keyword controls the treatment of user-missing values of nominal independent (predictor) variables.
MISSING. Handling of user-missing values of nominal predictor variables depends on the growing method. This is the default.
VALID. Treat user-missing values of nominal independent variables as valid values. User-missing values of nominal independent variables are treated as ordinary values in tree growing and classification.
Method-Dependent Rules
If some, but not all independent variable values are system-or user-missing:
- For CHAID and Exhaustive CHAID, system- and user-missing scale and ordinal independent variable values are included in the tree-growing process as a "floating" category that is allowed to merge with other categories in tree nodes. By default, nominal user-missing values are also handled in this fashion.
- For CRT and QUEST, cases with missing independent variable values are excluded from the tree-growing process but are classified using surrogate predictors, if surrogates are included in the method. By default, nominal user-missing values are also handled in this fashion.
Weight Variables
If the analysis uses frequency weights or an influence variable, all cases with zero, negative, system-missing, or user-missing values of either variable are excluded from the analysis.