Creating an empirical typology: A review of cluster analysis and other classification techniques

Abstract

As computers have become more powerful, researchers have begun to use more and more complex multivariate statistical procedures to understand data sets. Cluster analysis has flourished in the last decade, yet its complexity and reputation as a form of "soft" statistics continues to hamper its use in many ways. This article is intended as an introduction to both the science and the art of cluster analysis to the reader who has only a basic knowledge of the technique. In addition, a working example is provided to illustrate aspects of the process of completing a cluster analysis.