ERROR: "Tree Ensemble Learner" and "Random Forest Learner" show incorrect counts in Tree View

When using "Tree Ensemble Learner" or "Random Forest Learner" and you right click the module and select "Views: Tree Views" to look at the nodes in the model. Notice that all of the nodes in the tree show a incorrect "n" count of 1.

Even when you know the models is using 1000's of data points, and even when you specify that a child cannot have a count of 1.

This makes it very difficult to read the model to tell where the data is. I can't tell the difference of big groups and small groups.

Hello meltz,

this is actually intended behaviour (kind of). A random forest is a very memory intensive model and therefore we only store those counts if the user explicitly tells the node to do so.

To achieve this, check the checkbox "Save target distribution in tree nodes (memory expensive - only important for tree view and PMML export)" which can be found in the Options tab under Misc Options for the Random Forest Learner and in the Attribute Selection tab for the Tree Ensemble Learner.

If this doesn't solve the issue, please let me know.

Sorry for the inconvenience.

Cheers,

nemad