Formal Inference-based Recursive Modelling (FIRM)
I am trying to rebuild an algorithm called "ormal Inference-based Recursive Modelling Analysis " (FIRM) in RapidMiner, as explained here:
As far as I Understand it is a iterative search using Decision trees to select "good" attributes as judged by the Performance Operator.
Is Loop Feature Subset therefor the best mothod to build such a workflow? Is there anything I have be take care of, for example setting of the minumum information gain etc.?