Simplification of selections

[lionelkusch]

I have a few simplications suggestions.

Originally posted by @bthirion in #361 (review)

[lionelkusch]

For importance_selection:

hidimstat/src/hidimstat/base_variable_importance.py

Line 153 in 894ff9d

threshold_max : float, default=None

I'm not sure whether this argument really makes sense ?
I think I would have a unique threshold argument for this function.

Originally posted by @bthirion in #361 (comment)

For pvalue_selection:

hidimstat/src/hidimstat/base_variable_importance.py

Line 191 in 894ff9d

threshold_min : float, default=None

similarly, I don't see any use case for threshold_min here.

Originally posted by @bthirion in #361 (comment)

hidimstat/src/hidimstat/base_variable_importance.py

Line 193 in 894ff9d

alternative_hypothesis : bool, default=False

I don't see the use case for alternative hypothesis.

Originally posted by @bthirion in #361 (comment)

For fdr_selection:

hidimstat/src/hidimstat/base_variable_importance.py

Line 245 in 894ff9d

alternative_hippothesis: bool or None, default=False

Same thing here, I don't see any reason to consider an alternative hypothesis. This is because importance tests are all one-sided tests that test whether importance is greater 0 (=significantly different from 0, in that case).

Originally posted by @bthirion in #361 (comment)

[lionelkusch]

• I agree with the comments regarding simplifying the signature of selection functions.

Originally posted by @jpaillard in #361 (review)

[lionelkusch]

The option alternative_hippothesis has been introduced to keep the behaviour of EnCluDL.

The threshold_min and threshold_max have been added to it in the two selection functions for generality. The use case, I see, is a selection to discard some variables.

[bthirion]

Overall there is a semantic issue. EncluDL performs a signed statistical test, which is not strictly speaking an importance measure.