[pymvpa] GLMMapper

[Martin Sjøgård]

We are trying to get the GLM mapper to extract regressor betas for use in hyperalignment, but we're running into some problems and want to check if our way of constructing the mapper seems appropriate.


As you can see, we're using the GLMMapper with arguments ds and regs, as shown below:

--------------------------------------------------


ds = fmri_dataset(datafile, targets=myattributes.targets, chunks=myattributes.chunks, mask=roi)


pe = mvpa2.mappers.glm.GLMMapper(ds, regs)


print ds --> <Dataset: 1001x190059 at int16, <sa: chunks,targets,time_coords,time_indices>, <fa: voxel_indices>, <a: imgaffine,imghdr,imgtype,mapper,voxel_dim,voxel_eldim>>


print regs --> ['learn101', 'learn102', 'learn103', 'learn104', 'learn105', 'learn106', 'learn107', 'learn108', 'learn109', 'learn110', 'learn111', 'learn112', 'learn113', 'learn114', 'learn115', 'learn116', 'learn117', 'learn118', 'learn119', 'learn120', 'learn121', 'learn122', 'learn123', 'learn124', 'learn125', 'learn126', 'learn127', 'learn128', 'learn129', 'learn130', 'learn131', 'learn132', 'learn133', 'learn134', 'learn135', 'odd_even', 'fix']


print pe --> <GLMMapper>

--------------------------------------------------------------


So, our question is this: Is there a problem in the way we use the GLMMapper function? As you see, when we print ds we get a description of the dataset, but when we print pe we just get '<GLMMapper>'.



Martin Sjøgård
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-exppsy-pymvpa/attachments/20160310/a1c7eca5/attachment.html>