Confusion Matrix for a Nominal Target Variable in Enterprise Miner

[josh_winters]

I have a nominal target variable with 6 levels that I am builidng a classification model for.  In the out put of the model selection node the Event Classification Table (Confusion Matrix) looks like it is showing the results for just one of the levels of the target variable (see below for output)

Why is that and how do I get it to show the results for all levels of the target variable?? I am not a very good base sas programer by the way.  Also, it looks like performance measures are being calculated off the results for just this one level, is that true?

Event Classification Table
Model Selection based on Valid: Roc Index (_VAUR_)
 
                              Data             Target   False    True     False    True
Model Node  Model Description Role     Target  Label  Negative Negative Positive Positive
 
TextRule    Text Rule Builder TRAIN    Product            3       650       2        7
TextRule    Text Rule Builder VALIDATE Product            1       442       .        7
Tree        Decision Tree     TRAIN    Product            1       647       5        9
Tree        Decision Tree     VALIDATE Product            2       435       7        6
Boost       Gradient Boosting TRAIN    Product            3       652       0        7
Boost       Gradient Boosting VALIDATE Product            5       442       0        3
AutoNeural2 AutoNeural (2)    TRAIN    Product            1       651       1        9
AutoNeural2 AutoNeural (2)    VALIDATE Product            1       441       1        7
Ensmbl      Ensemble          TRAIN    Product            4       652       .        6
Ensmbl      Ensemble          VALIDATE Product            3       440       2        5

[WendyCzika · Accepted solution]

In the Results of the Model Comparison node, go to View>Assessment>Classification Chart and this displays the information that I think you need, and you can see the numbers by then selecting View>Table.  The information is displayed in a slightly different manner than below, but gives you the % correct and incorrect for each level of the target for each model.

Hope that helps!