Double Clustering: A Data Mining Methodology for Discovery of Causality
Andrzej Matuszewski · 2002
A generalization of the classical cluster analysis is proposed. From one point of view it is just a subsequent cluster analysis applied to the centers of clusters obtained in the first clustering. Such a double clustering procedure attains its effectiveness especially when a certain structure of causality (which often exists in the real, large data sets) is taken into account. Since we introduce a rather broad class of algorithms — which seems to be a quite general attempt — a very specific form of presentation is necessary. We present therefore methodology for a specific dataset to make the development of analysis and interpretation of its results more concrete.