In Oracle Data Miner if you are building decision trees (model building), why do we need to specify preferred target value in model options. Does it change the tree if I change the preferred variable? I tried for one data set with one target variable having cardinality of 2 but I get the same tree while choosing different preferred target value. Can anybody explain?
You must be using the Data Miner Classic UI (not Data Miner Workflow).
Data Miner Classic utilizes a Mining Activity to build and test models.
In the case of a Classification Mining Activity, you specify the build and test specifications.
The specific target value is for use in the Lift and ROC test result calculations.
In Data Miner Workflow the topn(100 or so, I forget the current setting) distinct target values have their lift results generated.