Clustering with noise and MDS
Clustering with noise uses a set of differences as input
(the original dialect differences) and produces a new set of differences
(the cophenetic differences). We can apply MDS on these new differences, just
like we did in normal MDS.
This procedure results in maps with only four major distinctive colours, and
only small variations within each colour. To see more detail, you
have to zoom in on an area, and repeat the procedure.
Limitations of normal MDS apply.