The Multidimensional Scaling platform generates a plot of proximities among a set of objects. This plot can be used to visually explore structure in a data set. MDS is a multivariate technique that is used to visualize the patterns of proximities (distances, similarities) among a set of objects in a small number of dimensions. MDS is applied to a distance matrix. The coordinates for the MDS plot are obtained by minimizing a stress function (the difference between the actual and predicted proximities).
The term distance can refer to a measure of physical distance, such as between cities. More often distance is a subjective assessment rather than a precise measurement. Proximities can measure perceived similarities between brands of a product, correlations of crime rates, or economic similarities for a sample of countries. Distance can also be called proximity or similarity (dissimilarity). If the data are given as an attribute list, then a distance matrix is first constructed from the attribute list.
For more information about multidimensional scaling, see Borg and Groenen (2005) or Jackson (2003).