mvpa2.measures.noiseperturbation.NoisePerturbationSensitivity¶
- class mvpa2.measures.noiseperturbation.NoisePerturbationSensitivity(datameasure, noise=<built-in method normal of mtrand.RandomState object at 0x2b07fbdd86a8>)¶
Sensitivity based on the effect of noise perturbation on a measure.
This is a FeaturewiseMeasure that uses a scalar Measure and selective noise perturbation to compute a sensitivity map.
First the scalar Measure computed using the original dataset. Next the data measure is computed multiple times each with a single feature in the dataset perturbed by noise. The resulting difference in the scalar Measure is used as the sensitivity for the respective perturbed feature. Large differences are treated as an indicator of a feature having great impact on the scalar Measure.
Notes
The computed sensitivity map might have positive and negative values!
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.
- training_time+: Time (in seconds) it took to train the learner
(Conditional attributes enabled by default suffixed with +)
Parameters : datameasure : Measure
Used to quantify the effect of noise perturbation.
noise: Callable :
Used to generate noise. The noise generator has to return an 1d array of n values when called the size=n keyword argument. This is the default interface of the random number generators in NumPy’s random module.
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
- is_trained = True¶
Indicate that this measure is always trained.
