rsatoolbox.rdm.transform module

transforms, which can be applied to RDMs

rsatoolbox.rdm.transform.positive_transform(rdms)

sets all negative entries in an RDM to zero and returns a new RDMs

Parameters:

rdms (RDMs) – RDMs object

Returns:

RDMs object with sqrt transformed dissimilarities

Return type:

rdms_new(RDMs)

rsatoolbox.rdm.transform.rank_transform(rdms: RDMs, method='average')

applies a rank_transform and generates a new RDMs object This assigns a rank to each dissimilarity estimate in the RDM, deals with rank ties and saves ranks as new dissimilarity estimates. As an effect, all non-diagonal entries of the RDM will range from 1 to (n_dim²-n_dim)/2, if the RDM has the dimensions n_dim x n_dim.

Parameters:

• rdms (RDMs) – RDMs object

• method (String) – controls how ranks are assigned to equal values options are: ‘average’, ‘min’, ‘max’, ‘dense’, ‘ordinal’

Returns:

RDMs object with rank transformed dissimilarities

Return type:

rdms_new(RDMs)

rsatoolbox.rdm.transform.sqrt_transform(rdms)

applies a square root transform and generates a new RDMs object This sets values blow 0 to 0 and takes a square root of each entry. It also adds a sqrt to the dissimilarity_measure entry.

Parameters:

rdms (RDMs) – RDMs object

Returns:

RDMs object with sqrt transformed dissimilarities

Return type:

rdms_new(RDMs)

rsatoolbox.rdm.transform.transform(rdms, fun)

applies an arbitray function fun to the dissimilarities and returns a new RDMs object.

Parameters:

rdms (RDMs) – RDMs object

Returns:

RDMs object with sqrt transformed dissimilarities

Return type:

rdms_new(RDMs)