<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
</head>
<body text="#000000" bgcolor="#FFFFFF">
Hi,<br>
<br>
sometimes I still encounter some problems with the transition from
mvpa to mvpa2. This time, I would like to run a Searchlight on an
artificial dataset. Previously that could be done like this:<br>
<br>
<meta charset="utf-8">
from sklearn.datasets import make_classification<br>
X,y = make_classification(n_samples=55, n_features=630, n_classes=2)<br>
ds = MaskedDataset(samples=X, labels=y, chunks=np.arange(y.size))<br>
cv = CrossValidatedTransferError(<br>
TransferError(LinearCSVMC()),<br>
OddEvenSplitter())<br>
sl = Searchlight(cv, radius=5)<br>
<br>
In mvpa2 I could do the same<br>
<br>
ds = Dataset(X)<br>
ds.targets = y<br>
ds.chunks = np.arange(55)<br>
<br>
but the sphere_searchlight requires a 'space' parameter. Here I
don't have real neighbor-space relations. It can be completely
random or neighbors in the np.array. How would you define that for a
synthetic dataset?<br>
<br>
Thanks in advance,<br>
Matthias<br>
</body>
</html>