Consider constructing a certain classifier in a situation with 500 features total. 50 of them are truly informative about the classes. Another 50 features are direct copies of the first 50 features. The remaining 400 features are not informative(i.e., noisy features). Assume there is sufficient samples to reliably assess how useful features are, and the feature selection methods are using good thresholds.

(1) How many features will be selected by mutual information filtering method? Explain it briefly in two or three sentences. (2) How many features will be selected by a wrapper method? Explain it briefly in two or three sentences.