ICSV.Server.MDS.MDS Class Reference

Easy access interface to the MDSJ component for Multi-Dimensional Scaling More...

Static Public Member Functions
static double[][]	Mds (out string report, double[][] D, int dimensions, int rounds=0, int timeout=0, int stresschange=0)
	Runs a MDS algorithm. More precisely an approximation using landmark points. See also V. de Silva, J.B. Tenenbaum, Sparse multidimensional scaling using landmark points. More...

Detailed Description

Easy access interface to the MDSJ component for Multi-Dimensional Scaling

MDSJ available at http://www.inf.uni-konstanz.de/algo/software/mdsj/ Documentation at http://www.inf.uni-konstanz.de/algo/software/mdsj/javadoc/

Member Function Documentation

static double [][] ICSV.Server.MDS.MDS.Mds ( out string  report, double  D[][], int  dimensions, int  rounds = 0, int  timeout = 0, int  stresschange = 0  )

static

Runs a MDS algorithm. More precisely an approximation using landmark points.

See also

V. de Silva, J.B. Tenenbaum, Sparse multidimensional scaling using landmark points.

Parameters

report	a documentation of the result's precision and remaining stress
D	the square-sized distance matrix used to compute positions
dimensions	the dimension of the resulting Euclidean space. Typically 2 or 3.
rounds	the maximum count of iterations used for computation. If the parameter is 0 it is ignored.
timeout	the maximum time allowed for computation. In milliseconds. If the parameter is 0 it is ignored.
stresschange	the relative stress change between iterations must be below 10^(-stresschange). If the parameter is 0 it is ignored.

Returns

the resulting positions in the Euclidean space. Size: D x N, where D is the number of dimensions and N is the number of elements in the distance matrix (i.e. number of rows or columns)