[pymvpa] get_sensitivity_analyzer for PLR

[Yaroslav Halchenko]

sorry about the delay ... you have hit one of the corners where we
should clean up our interface/functionality to make such obvious things
readily available without needing to code anything

Meanwhile I have pushed the fix -- PLR now exposes its "sensitivity"
(linear weights).  available from git , see

http://www.pymvpa.org/installation.html#building-from-source

Cheers,

On Sat, 24 Dec 2011, Hiroyuki AKAMA wrote:

> Dear PyMVPA developpers

> I am using SMLR as well as PLR, and 'get_sensitivity_analyzer' perfectly worked with SMLR but not with PLR. 

> <<Error Message>>
>  File "PLRAvVolsSensitivity.py", line 158, in <module>
>     weightAnalyser = anovaSelectedPLR.get_sensitivity_analyzer(postproc=M.maxofabs_sample())
>   File "/usr/lib/pymodules/python2.6/mvpa2/misc/args.py", line 73, in do_group_kwargs
>     return method(self, *args_, **kwargs__)
>   File "/usr/lib/pymodules/python2.6/mvpa2/clfs/meta.py", line 330, in get_sensitivity_analyzer
>     analyzer=self.__clf.get_sensitivity_analyzer(**slave_kwargs),
>   File "/usr/lib/pymodules/python2.6/mvpa2/clfs/base.py", line 491, in get_sensitivity_analyzer
>     raise NotImplementedError
> NotImplementedError

> We think get_sensitivity_analyzer just hasn't coded that yet. 

> What should we do for the time being? Is it soluble adding lines to some py files?

> FYI.
> anovaSelectedPLR = M.FeatureSelectionClassifier(
> 	M.PLR(),
> 	M.SensitivityBasedFeatureSelection(
> 		M.OneWayAnova(),
> 		M.FixedNElementTailSelector(500, mode='select', tail='upper')
> 	),
> )

-- 
=------------------------------------------------------------------=
Keep in touch                                     www.onerussian.com
Yaroslav Halchenko                 www.ohloh.net/accounts/yarikoptic