[pymvpa] Sensitivity analysis

[J.A. Etzel]

I'd add

Lee, S., Halder, S., Kübler, A., Birbaumer, N., Sitaram, R., 2010.
Effective functional mapping of fMRI data with support-vector machines.
Human Brain Mapping 31, 1502-1511.

to the must-read list for interpreting weight parameters.

I think this is a trickier topic than sometimes appreciated (even for 
linear svms), given the interaction of weights and activation values, 
the high correlation between voxels, and typical low stability (e.g. 
massive differences in weight vectors between cross-validation folds or 
if a single trial is dropped).

Jo


On 6/14/2011 3:59 PM, Roberto Guidotti wrote:
> 2011/6/14 Yaroslav Halchenko <debian at onerussian.com
> <mailto:debian at onerussian.com>>