[pymvpa] classification selection

[Liang, Guangsheng]

Oh sorry. I forgot to post that info. Please see below please.
Thanks!


"Dataset: 1770x131072 at float64, <sa: chunks,targets,time_coords,time_indices>, <fa: voxel_indices>, <a: imgaffine,imghdr,imgtype,mapper,voxel_dim,voxel_eldim>
stats: mean=-2.65305e-18 std=0.99975 var=0.999501 min=-10.8893 max=18.7883

Counts of targets in each chunk:
  chunks\\targets post_rest pre_rest
                      ---       ---
         0.0          177       177
         1.0          177       177
         2.0          177       177
         3.0          177       177
         4.0          177       177

  Summary for targets across chunks
     targets mean std min max #chunks
   post_rest  177  0  177 177    5
   pre_rest  177  0  177 177    5

  Summary for chunks across targets
    chunks mean std min max #targets
      0     177  0  177 177     2
      1     177  0  177 177     2
      2     177  0  177 177     2
      3     177  0  177 177     2
      4     177  0  177 177     2

  Sequence statistics for 1770 entries from set ['post_rest', 'pre_rest']\nCounter-balance table for orders up to 2:
     Targets/Order  O1      |   O2      |
       post_rest:  884  0   |  883  0   |
       pre_rest:    1  884  |   2  883  |
                                                
  Correlations: min=-1 max=1 mean=-0.00057 sum(abs)=8.8e+02"





On 6/14/16, 5:24 PM, "Pkg-ExpPsy-PyMVPA on behalf of Yaroslav Halchenko" <pkg-exppsy-pymvpa-bounces+guangsheng.liang=ttu.edu at lists.alioth.debian.org on behalf of debian at onerussian.com> wrote:

>
>On Tue, 14 Jun 2016, Liang, Guangsheng wrote:
>
>>    Hello PyMVPA community,
>
>>    I just start using PyMVPA for some fMRI data analysis, and I got a
>>    question about classification selection.
>
>>    I firstly used LinearCSVMC as my initial classifier, but it ran very very
>>    slow (it kept running for more than 2 weeks).
>
>>    I am not quite sure where did I go wrong.  I then used another classifier
>>    - kNN. It ran much faster than the LinearCSVMC.
>
>>    I was wondering if there was any restriction or cautions on using kNN for
>>    fMRI data analysis?
>
>it all depends on details of your study...   at least having a clue
>(print your_ds.summary()) could be of help.
>
>-- 
>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        
>
>_______________________________________________
>Pkg-ExpPsy-PyMVPA mailing list
>Pkg-ExpPsy-PyMVPA at lists.alioth.debian.org
>http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-exppsy-pymvpa