This paper gives a description of (reverse-) clustering-methods. A very useful technique in reducing the complexity of multi causal relations, and for re-arranging the variables in a database. Emphasis is put on the basics aspects of the technique, in terminology, some recent developments and its application. Also several general research-aspects are discussed: like checking for a small data-dispersion before drawing any conclusions, selecting those variables which are (time and/or place) independent, and so on.
Abstract