Explain Model Action Set

When you specify the generatedOut parameter, the linearExplainer action produces an output CAS table that contains the rows of data that have been scored by the machine learning model before the training of the regression model. You can also use this table in future runs of the linearExplainer action by specifying the table in the table parameter and also by specifying NONE in the method subparameter of the dataGeneration parameter. This enables you to use the same generated data for multiple analyses.

You can also use this table for custom analysis outside the linearExplainer action.

The output CAS table contains the generated variable _kernelWeight_. If you specify ALL or NOMINALS for the binaryEncoding parameter, the linearExplainer action also generates variables starting with _bCode_. Using an input data table that contains these variable names might cause name collisions.