Add an axis to a dataset at an arbitrary position.
This mapper can be useful when there is need for aggregating multiple datasets, where it is often necessary or at least useful to have a dedicated aggregation axis. An axis can be added at any position
When adding an axis that causes the current sample (1st) or feature axis (2nd) to shift the corresponding attribute collections are modified to accomodate the change. This typically means also adding an axis at the corresponding position of the attribute arrays. A special case is, however, prepending an axis to the dataset, i.e. shifting both sample and feature axis towards the back. In this case all feature attibutes are duplicated to match the new number of features (formaly the number of samples).
Notes
Available conditional attributes:
(Conditional attributes enabled by default suffixed with +)
Examples
>>> from mvpa2.datasets.base import Dataset
>>> from mvpa2.mappers.shape import AddAxisMapper
>>> ds = Dataset(np.arange(24).reshape(2,3,4))
>>> am = AddAxisMapper(pos=1)
>>> print am(ds).shape
(2, 1, 3, 4)
Parameters : | pos : int
enable_ca : None or list of str
disable_ca : None or list of str
auto_train : bool
force_train : bool
space: str, optional :
postproc : Node instance, optional
descr : str
|
---|