mvpa2.measures.base.TransferMeasure

Inheritance diagram of TransferMeasure

class mvpa2.measures.base.TransferMeasure(measure, splitter, **kwargs)

Train and run a measure on two different parts of a dataset.

Upon calling a TransferMeasure instance with a dataset the input dataset is passed to a Splitter to will generate dataset subsets. The first generated dataset is used to train an arbitray embedded `Measure. Once trained, the measure is then called with the second generated dataset and the result is returned.

Notes

Available conditional attributes:

  • calling_time+: Time (in seconds) it took to call the node
  • null_prob+: None
  • null_t: None
  • raw_results: Computed results before invoking postproc. Stored only if postproc is not None.
  • stats: Optional summary statistics about the transfer performance
  • training_stats: Summary statistics about the training status of the learner
  • training_time+: Time (in seconds) it took to train the learner

(Conditional attributes enabled by default suffixed with +)

Methods

generate(ds) Yield processing results.
get_postproc() Returns the post-processing node or None.
get_space() Query the processing space name of this node.
reset()
set_postproc(node) Assigns a post-processing node Set to None to disable postprocessing.
set_space(name) Set the processing space name of this node.
train(ds) The default implementation calls _pretrain(), _train(), and finally _posttrain().
untrain() Reverts changes in the state of this node caused by previous training
Parameters:

measure: Measure :

This measure instance is trained on the first dataset and called with the second.

splitter: Splitter :

This splitter instance has to generate at least two dataset splits when called with the input dataset. The first split is used to train the measure, the second split is used to run the trained measure.

enable_ca : None or list of str

Names of the conditional attributes which should be enabled in addition to the default ones

disable_ca : None or list of str

Names of the conditional attributes which should be disabled

null_dist : instance of distribution estimator

The estimated distribution is used to assign a probability for a certain value of the computed measure.

auto_train : bool

Flag whether the learner will automatically train itself on the input dataset when called untrained.

force_train : bool

Flag whether the learner will enforce training on the input dataset upon every call.

space: str, optional :

Name of the ‘processing space’. The actual meaning of this argument heavily depends on the sub-class implementation. In general, this is a trigger that tells the node to compute and store information about the input data that is “interesting” in the context of the corresponding processing in the output dataset.

postproc : Node instance, optional

Node to perform post-processing of results. This node is applied in __call__() to perform a final processing step on the to be result dataset. If None, nothing is done.

descr : str

Description of the instance

Methods

generate(ds) Yield processing results.
get_postproc() Returns the post-processing node or None.
get_space() Query the processing space name of this node.
reset()
set_postproc(node) Assigns a post-processing node Set to None to disable postprocessing.
set_space(name) Set the processing space name of this node.
train(ds) The default implementation calls _pretrain(), _train(), and finally _posttrain().
untrain() Reverts changes in the state of this node caused by previous training
is_trained = True

Indicate that this measure is always trained.

measure
splitter

NeuroDebian

NITRC-listed