mvpa2.measures.winner.mean_group_sample(attrs, attrfx='merge', **kwargs)

Returns a mapper that computes the mean samples of unique sample groups.

The sample groups are identified by the unique combination of all values of a set of provided sample attributes. Order of output samples might differ from original and correspond to sorted order of corresponding attrs by default. Use order='occurrence' if you would like to maintain the order.


attrs : list

List of sample attributes whose unique values will be used to identify the samples groups.

attrfx : ‘merge’ or callable, optional

Callable that is used to determine the sample attributes of the computed mean samples. By default this will be a string representation of all unique value of a particular attribute in any sample group. If there is only a single value in a group it will be used as the new attribute value.


FxMapper instance.