NLFISHERM

Non-linear Fisher Mapping according to Marco Loog

    W = NLFISHERM(A,N)

Input
A	Dataset
N	Number of dimensions (optional; default: MIN(K,C)-1, where
K	is the dimensionality of A and C is the number of classes)

Output
W	Non-linear Fisher mapping

Description

Finds a mapping of the labeled dataset A to a N-dimensional linear  subspace emphasizing the class separability for neighboring classes.

Reference(s)

1. R. Duin, M. Loog and R. Haeb-Umbach, Multi-Class Linear Feature Extraction by Nonlinear PCA, in: ICPR15, 15th Int. Conf. on Pattern Recognition, vol.2, IEEE Computer Society Press, 2000, 398-401.
2. M. Loog, R.P.W. Duin and R. Haeb-Umbach, Multiclass Linear Dimension Reduction by Weighted Pairwise Fisher Criteria, IEEE Trans. on Pattern Analysis and Machine Intelligence, vol.23, no.7, 2001, 762-766.

See also

mappings, datasets, fisherm, klm, pca,