The documentation for the FeatureExtractor option for the Classfiy[] and Predict[] functions only provides the vaguest description of what feature selection is performed under this option, i.e.

FeatureExtractor "Minimal"] indicates that the internal preprocessing should be as simple as possible

I would like to know what is meant by "as simple as possible internal processing" - does it mean that there is no feature selection at all, or, alternatively, if some type of feature selection is performed, what method is involved? Ideally, I would like to disable feature selection altogether for Classify and Predict so I can compare the performance of various classifier algorithms with and without feature selection, is this possible in the framework of these functions?

Thank you for any additional information.