<html>
<body>
Thanks Michael,<br><br>
At 10:15 AM 10/1/2012, <b>Michael Hanke</b> wrote:<br><br>
<br><br>
<blockquote type=cite class=cite cite=""><pre>> Errors:
> 'gcc' is not recognized as an internal or external command,
> operable program or batch file.
> C:\Python27\lib\site-packages\scipy\integrate\quadpack.py:288:
You don't have built pymvpa's extensions properly. How did you
install
it.</pre><font face="Courier New, Courier"></font></blockquote><br>
I got it from Christoph Gohlke's excellent build bot:
<a href="http://www.lfd.uci.edu/~gohlke/pythonlibs/" eudora="autourl">
http://www.lfd.uci.edu/~gohlke/pythonlibs/<br>
</a>He did have a ticket for msvc compile errors earlier:
<a href="https://github.com/PyMVPA/PyMVPA/pull/58" eudora="autourl">
https://github.com/PyMVPA/PyMVPA/pull/58<br>
</a>and might be amenable to debugging this build if we can help him in
the right direction.<br>
Incidentally, I added gcc on this system, and the only errors are
now:<br>
<font face="MS Shell Dlg 2, Geneva" size=1 color="#400000">
C:\Python27\lib\site-packages\scipy\integrate\quadpack.py:288:
UserWarning: Extremely bad integrand behavior occurs at some points of
the<br>
integration interval.<br>
warnings.warn(msg)<br>
C:\Python27\lib\site-packages\mvpa2\misc\errorfx.py:106: RuntimeWarning:
invalid value encountered in divide<br>
([0], np.cumsum(~t)/(~t).sum(dtype=np.float), [1]))<br>
C:\Python27\lib\site-packages\scipy\stats\stats.py:274: RuntimeWarning:
invalid value encountered in double_scalars<br>
return np.mean(x,axis)/factor<br>
C:\Python27\lib\site-packages\mvpa2\misc\errorfx.py:102: RuntimeWarning:
invalid value encountered in divide<br>
([0], np.cumsum(t)/t.sum(dtype=np.float), [1]))<br><br>
<br>
</font><blockquote type=cite class=cite cite=""><pre>Here is also the
usual advice: "Don't do this at home!" On
windows
it is a lot more convenient to get the NeuroDebian virtual machine and
install PyMVPA inside -- shouldn't take you more than 5 min to set up
</pre><font face="Courier New, Courier"></font></blockquote><br>
Downloading now for myself; the issue there is that I write code for a
small group of Windows7 users here, and I'd like to incorporate mvpa with
the usual Python exes I make for them. The 2nd best alternative is to run
it on Amazon's *NIX cloud service we use.<br>
I'll also try the py2.6 version next.<br><br>
<br><br>
<blockquote type=cite class=cite cite=""><pre>.... <SNIP>
stats['FP'] /
(1.0*stats["P'"])
> C:\Python27\lib\site-packages\mvpa2\measures\anova.py:111:
> RuntimeWarning: invalid value encountered in divide
> msb = ssbn / float(dfbn)
>
Not sure why this happens -- could be specific to you input dataset.
</pre><font face="Courier New, Courier"></font></blockquote><br>
attached, with a revised script<br>
The runtime errors and the mvpa2 warning
"<font face="MS Shell Dlg 2, Geneva" size=1> Only 1 sets have
estimates assigned from 2
sets</font><font face="MS Shell Dlg 2, Geneva">"</font> looks like
my data setup (?)<br><br>
<br><br>
<blockquote type=cite class=cite cite=""><pre>> WARNING: Obtained
degenerate data with zero norm for training
of
> <LinearCSVMC>. Scaling of C cannot be done.
There seem to be a problem with the dataset. do you have invariant
features in it?
</pre><font face="Courier New, Courier"></font></blockquote><br>
I don't belive so, the features are FFT peaks' power and f.<br>
I added <br>
mvpa2.datasets.miscfx.coarsen_chunks(ds,
nchunks=4)<br>
mvpa2.datasets.miscfx.remove_invariant_features(ds)<br>
to the code, with no apparent change.<br><br>
I do note that CART (tested prior on this set) will _always_ split data
even on junk, there is no guarantee that there are any useful
correlations in the data presented...<br>
I'm looking at mlpy right now as well.<br><br>
Best,<br>
Ray<br>
</body>
</html>