From Complexity to Simplicity: On the Application of Three Techniques for Multivariate Data Analysis

The aim of this paper is to give an overview of three important techniques; factor, cluster and discriminant analysis. I find it necessary and more important to have an understanding of the basic assumptions and the underlying foundations of the methods rather than a thorough mathematical understanding of the algorithms. I will therefore account for a description of these methods based on a geometrical viewpoint and a more conceptual view rather than mathematical. A general description of the three techniques will be followed by a case, showing an application of all three techniques in the same study.