Package mds provides multidimensional scaling functions.
TorgersonScaling converts a dissimilarity matrix to a matrix containing Euclidean coordinates. TorgersonScaling places the coordinates in dst and returns it and the number of positive Eigenvalues if successful. If the scaling is not successful, dst is returned, but will not be a valid scaling. When the scaling is successful, mds will be resized to k columns wide. Eigenvalues will be copied into eigdst and returned as eig if it is provided.
If dst is nil, a new mat.Dense is allocated. If dst is not a zero matrix, the dimensions of dst and dis must match otherwise TorgersonScaling will panic. The dis matrix must be square or TorgersonScaling will panic.