<html>
<head>
<style><!--
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Tahoma
}
--></style></head>
<body class='hmmessage'><div dir='ltr'>
Hi,<BR> <BR>A quick question about how the permutation is done for the following code from the permutation example in pymva tutorial:<BR> <BR>=======<BR>clf = LinearCSVMC()<br>partitioner = NFoldPartitioner()<br>permutator = AttributePermutator('targets', count=200)<br>distr_est = MCNullDist(permutator, tail='left', enable_ca=['dist_samples'])<br>cv = CrossValidation(clf, partitioner,<br> errorfx=mean_mismatch_error,<br> postproc=mean_sample(),<br> null_dist=distr_est,<br> enable_ca=['stats'])<BR>=======<BR> <BR>Is the permutation done within each chunk or across all chunks? If it is across all chunks, is the number of trials for each target still the same within each chunk after each permutation? I imagine that the labels for both target and chunk would be permutated together so that the number of trials would remain the same for all targets?<BR> <BR>Also, according to the webpage, the above codes permute both testing and training datasets, right?<BR> <BR>Many thanks in advance!<BR> <BR>Meng<BR> </div></body>
</html>