Multi-Dimensional Scaling (MDS)

Multi-Dimensional Scaling (MDS), sometimes also called Principal Coordinates Analysis (PCoA), is a non-hierarchic grouping method. Rather than starting from the data set as Principal Components Analysis (PCA) does, MDS uses the similarity matrix as input, which has the advantage over PCA that it can be applied directly to pairwise-compared banding patterns.

3D representation of an MDS

For MDS, the same three-dimensional graphs and viewing tools are available as for PCA and discriminant analysis.