<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=gb2312">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<div style="font-family: Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<pre>Dear all,
I am running hyperalignment between-subject analysis. Before calling Hyperalignment(), I stacked the </pre>
<pre>datasets across the subjects with all runs and did Anova-based feature selection. Hyperalignment() fails </pre>
<pre>for the dataset with the error message: "For now do not handle invariant in time datasets".</pre>
<pre><pre><br></pre><pre>It would be great if somebody has any kind of information about what causes this issue, and ways to </pre><pre>solve it. The problem is not fixed after I removed invariant features by using remove_invariant_features()</pre><pre>or set the svd in ProcrusteanMapper() to 'dgesvd'. </pre><pre><br></pre><pre>My codes are as follow:</pre><pre><span># split in training and testing sets<br></span><div>ds_train = dscon[np.array([r != test_run for r in dscon.sa.chunks], dtype='bool'),:]<br></div><div>ds_test = dscon[np.array([r == test_run for r in dscon.sa.chunks], dtype='bool'),:]<br></div><div>ds_train = remove_invariant_features(ds_train)<br></div><div><br></div><div># manual feature selection for every individual dataset in the list<br></div><div>anova = OneWayAnova(space='reg')<br></div><div>fscores = [anova(sd) for sd in ds_train]<br></div><div>eatsels = [StaticFeatureSelection(feature_selector(fscore)) for fscore in fscores]<br></div><div>ds_train_fs = [fs.forward(sd) for fs, sd in zip(featsels, ds_train)]<br></div><div> <br></div><div># Perform hyperalignment on the training data<br></div><div>hyper = Hyperalignment(alignment=ProcrusteanMapper(sspace='reg'))<br></div><span>hypmaps = hyper(ds_train_fs)</span><br></pre><pre><br></pre><pre>The error message is:</pre><pre><span>Traceback (most recent call last):<br></span><div><br></div><div> File "<ipython-input-377-74fee538844a>", line 2, in <module><br></div><div> hypmaps = hyper(ds_train_fs)<br></div><div><br></div><div> File "/usr/local/Anaconda/lib/python2.7/site-packages/mvpa2/algorithms/hyperalignment.py", line 339, in __call__<br></div><div> self.train(datasets)<br></div><div><br></div><div> File "/usr/local/Anaconda/lib/python2.7/site-packages/mvpa2/algorithms/hyperalignment.py", line 313, in train<br></div><div> residuals)<br></div><div><br></div><div> File "/usr/local/Anaconda/lib/python2.7/site-packages/mvpa2/algorithms/hyperalignment.py", line 419, in _level1<br></div><div> m.train(ds_new)<br></div><div><br></div><div> File "/usr/local/Anaconda/lib/python2.7/site-packages/mvpa2/base/learner.py", line 137, in train<br></div><div> self._train(ds)<br></div><div><br></div><div> File "/usr/local/Anaconda/lib/python2.7/site-packages/mvpa2/mappers/procrustean.py", line 123, in _train<br></div><div> raise ValueError, "For now do not handle invariant in time datasets"<br></div><div><br></div><div>ValueError: For now do not handle invariant in time datasets<br></div><span></span><br></pre><br></pre>
<pre>
</pre>
</div>
</body>
</html>