FeaturewiseMeasure that assesses feature stability across runs for each unique label by correlating label activity for pairwise combinations of the chunks.
If there are multiple samples with the same label in a single chunk (as is typically the case) this algorithm will take the featurewise average of the sample activations to get a single value per label, per chunk.
Notes
Available conditional attributes:
(Conditional attributes enabled by default suffixed with +)
Initialize
Parameters : | attr : str
enable_ca : None or list of str
disable_ca : None or list of str
null_dist : instance of distribution estimator
auto_train : bool
force_train : bool
space: str, optional :
postproc : Node instance, optional
descr : str
|
---|