mvpa2.base.dataset.split_by_feature_attribute

mvpa2.base.dataset.split_by_feature_attribute(ds, fa_label, raise_unequal_count=True)

Splits a dataset based on unique values of a feature attribute

Parameters:

d: Dataset :

input dataset

sa_label: str or list of str :

sample attribute label(s) on which the split is based

Returns:

ds: list of Dataset :

List with n datasets, if d.fa[fa_label] has n unique values