ABSTRACT
This chapter describes the two classes of techniques are multidimensional scaling and correspondence analysis. Multidimensional scaling and correspondence analysis both aim to help in understanding particular types of data by displaying the data graphically. Multidimensional scaling applied to proximity matrices is often useful in uncovering the dimensions on which similarity judgments are made. Classical multidimensional scaling seeks to represent a proximity matrix by a simple geometrical model or map. For more information of the connections between correspondence analysis and other multivariate methods, together with more general biplot techniques to visualize the results of these methods, see Greenacre. In some psychological work and in market research, proximity matrices arise from asking human subjects to make judgments about the similarity or dissimilarity of objects or stimuli of interest. The corresponding advantages vary from being outdoors, having freedom, through family and human aspects to salaries, social advantages, schedules, and security.
