[pymvpa] PyMVPA WARNING messages

[Yaroslav Halchenko]

On Fri, 11 Mar 2016, Salim Al-wasity wrote:

> DearsI am quite new user of PyMVPA package and I have some warning messages:
> 1- When I use (from mvpa2.suite import *) I get this:
> Warning (from warnings module):  File "C:\Python27\lib\site-packages\numpy\lib\utils.py", line 99    warnings.warn(depdoc, DeprecationWarning)DeprecationWarning: `scipy.weave` is deprecated, use `weave` instead!
> Warning (from warnings module):  File "C:\Python27\lib\site-packages\scipy\integrate\quadpack.py", line 352    warnings.warn(msg, IntegrationWarning)IntegrationWarning: Extremely bad integrand behavior occurs at some points of the  integration interval.

not sure where 2nd one comes from atm.... but I guess both are safe to
ignore for now... we should address them eventually

> 2- When I train a classifier using a searchlight results:clf=LinearCSVMC()cv = CrossValidation(clf, NFoldPartitioner())sl = sphere_searchlight(cv, radius=3, postproc=mean_sample())sel=FractionTailSelector(0.005, mode='select', tail='lower')
> new_ds=ds[:,sel(res)]
> cvte=CrossValidation(clf, NFoldPartitioner(), enable_ca=['stats'])results = cvte(new_ds) ###print np.round(cvte.ca.stats.stats['ACC%'], 1)I get this error when I execute the last (print ) command:
> Warning (from warnings module):  File "C:\Python27\lib\site-packages\numpy\lib\utils.py", line 99    warnings.warn(depdoc, DeprecationWarning)DeprecationWarning: `nanmean` is deprecated!scipy.stats.nanmean is deprecated in scipy 0.15.0 in favour of numpy.nanmean.

also should be safe to ignore since didn't crash... we should address it

> 3- When I use (evds = fit_event_hrf_model(ds,events,time_attr='time_coords',condition_attr=('targets', 'chunks'))) I get this:
> Warning (from warnings module):  File "C:\Python27\lib\site-packages\nipy\modalities\fmri\design_matrix.py", line 229    warn('Matrix is singular at working precision, regularizing...')UserWarning: Matrix is singular at working precision, regularizing...
> Warning (from warnings module):  File "C:\Python27\lib\site-packages\nipy\modalities\fmri\glm.py", line 127    (ols_result.resid ** 2).sum(0))RuntimeWarning: invalid value encountered in divide

suggests that most probably design matrix is degenerate
(collinearities).  probably better to not ignore

-- 
Yaroslav O. Halchenko
Center for Open Neuroscience     http://centerforopenneuroscience.org
Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
Phone: +1 (603) 646-9834                       Fax: +1 (603) 646-1419
WWW:   http://www.linkedin.com/in/yarik