Troubleshooting
Problem
In AnswerTree, I wished to have the case weights reflected in the node statistics i.e. the counts and percents for categories of the target variable. These statistics do reflect frequency weights, but not case weights. My case weights were noninteger, so designating them as frequency weights would lead to their being rounded at the case level. To accomplish my goal, I had to apply the SPSS assignment rules in an SPSS syntax window to compute the node numbers as new variables in the SPSS file. I then designated the case weights as the WEIGHT variable in SPSS and used the TABLES procedure to produce a cross-tabulation of the target variable and the terminal node number. Do case weights and frequency weights operate in the SPSS Decision Tree procedure (TREE command) in the same manner as they do in AnswerTree? Is there an easier way to force node statistics to reflect the case weights?
Log InLog in to view more of this document
Historical Number
48684
Was this topic helpful?
Document Information
Modified date:
16 April 2020
UID
swg21480512