MDS_CS

MDS_CS Trainable mapping for classical scaling

    W = MDS_CS(D,N)
    W = D*MDS_CS([],N)
    W = D*MDS_CS(N)

Input
D	Square dissimilarity matrix of the size M x M
N	Desired output dimensionality (optional; default: 2)

Output
W	Classical scaling mapping

Description

A linear mapping of objects given by a symmetric distance matrix D with  a zero diagonal onto an N-dimensional Euclidean space such that the square  distances are preserved as much as possible.

D is assumed to approximate the Euclidean distances, i.e.  D_{ij} = sqrt(sum_k (x_i-x_j)^2).

Reference(s)

1. I. Borg and P. Groenen, Modern Multidimensional Scaling, Springer Verlag, Berlin, 1997.
2. T.F. Cox and M.A.A. Cox, Multidimensional Scaling, Chapman and Hall, London, 1994.

See also

mappings, mds_init, mds_cs,