[pymvpa] SensitivityBasedFeatureSelection for LinearCSVMC

[patrik andersson]

Hi,

I have a basic question about the sensitivity analysis for the
LinearCSVMC classifier.
Namely: What does it do?
Simple hey?! :)

Lets say I have;

clfr = LinearCSVMC(probability=1,C=1)
fsel = SensitivityBasedFeatureSelection(
			   clfr.get_sensitivity_analyzer(postproc=maxofabs_sample()),
			   FixedNElementTailSelector(nr_features_keep, mode='select', tail='upper'),
			   enable_ca = ['selected_ids'],
			   train_analyzer=1)

What comes out of the  clfr.get_sensitivity_analyzer function?
Is there some iteration or data folding going on inside?
What happens if fsel is called with multiclass data?

Answer this and I wish you an even happier x-mas ;)

P