multi-dimensional scaling

Page numbers are for draft copy at present; they will be replaced with correct numbers when final book is formatted. Chapter numbers are correct and will not change now.

Multi-dimensional scaling (MDS) creates a two (usually) dimensioanl represention of data that preserves closeness. Its input is a similarity matrix (S_ik) or a dissimilarity matrix (D_ik) between items an then allocates for each item i coordinates (x_i,y_i) such that, on average, when the Euclidean distance between (x_i,y_i) and (x_j,y_j) is small then S_ik) is large (or D_ik) is small) and vice versa.

Used on Chap. 8: page 160

Also known as MDS