COPULAM

Copula mapping

    [W,V] = COPULAM(A,K)
    [W,V] = A*COPULAM
    C = B*W

Input
A	Dataset, used for training the mapping
B	Dataset, same dimensionality as A, to be mapped
K	Desired output dimensionality, default dimensionality of A.

Output
W	Trained mapping
V	Trained inverse mapping
C	Copula transformed dataset

Description

The copula of a dataset transforms for every dimension every feature  value into the value of its cumulative feature density. So, A*W executes  the transform on the original data. A*W*V returns A again. This mapping  uses MAPSD in order to apply it on another dataset B of the same  dimensionality. It thereby creates an approximation trained by A.

If K is given, first a PCA to K dimensions is performed.

See also

datasets, mappings, pcam, mds, tsnem, mapsd,