This content refers to the previous stable release of PyMVPA. Please visit www.pymvpa.org for the most recent version of PyMVPA and its documentation.

misc.errorfx

Module: misc.errorfx

Inheritance diagram for mvpa.misc.errorfx:

Error functions helpers.

PyMVPA can use arbitrary function which takes 2 arguments: predictions and targets and spits out a scalar value. Functions below are for the convinience, and they confirm the agreement that ‘smaller’ is ‘better’

Classes

AUCErrorFx

class mvpa.misc.errorfx.AUCErrorFx

Bases: mvpa.misc.errorfx._ErrorFx

Computes the area under the ROC for the given the target and predicted to make the prediction.

MeanMismatchErrorFx

class mvpa.misc.errorfx.MeanMismatchErrorFx

Bases: mvpa.misc.errorfx._ErrorFx

Computes the percentage of mismatches between some target and some predicted values.

RMSErrorFx

class mvpa.misc.errorfx.RMSErrorFx

Bases: mvpa.misc.errorfx._ErrorFx

Computes the root mean squared error of some target and some predicted values.

RelativeRMSErrorFx

class mvpa.misc.errorfx.RelativeRMSErrorFx

Bases: mvpa.misc.errorfx._ErrorFx

Ratio between RMSE and root mean power of target output.

So it can be considered as a scaled RMSE – perfect reconstruction has values near 0, while no reconstruction has values around 1.0. Word of caution – it is not commutative, ie exchange of predicted and target might lead to completely different answers

Variance1SVFx

class mvpa.misc.errorfx.Variance1SVFx

Bases: mvpa.misc.errorfx._ErrorFx

Ratio of variance described by the first singular value component.

Of limited use – left for the sake of not wasting it

Functions

mvpa.misc.errorfx.meanPowerFx(data)

Returns mean power

Similar to var but without demeaning

mvpa.misc.errorfx.rootMeanPowerFx(data)

Returns root mean power

to be comparable against RMSE