Y = cmdscale(D)
Y = cmdscale(D, p)
[Y, e] = cmdscale(...)
cmdscale computes a classical multidimensional scaling configuration from a distance, dissimilarity, or similarity matrix.
D can be a square matrix or a distance vector accepted by squareform. If p is specified, only coordinates associated with positive eigenvalues among the first p dimensions are returned.
The vector e contains the ordered eigenvalues of the centered inner-product matrix. When p is specified, e contains at most p values.
X = [0 0; 1 0; 0 2; 2 2];
D = squareform(pdist(X));
[Y, e] = cmdscale(D)| Version | Description |
|---|---|
| 2.0.0 | initial version |