Evolutionary Rule Based Clustering for Making Fuzzy Object Oriented Database Models

This paper proposes a database clustering algorithm using genetic network programming (GNP) with the advantages of fuzzy object oriented database modeling. GNP creates clusters based on pattern classification, where a cluster label is assigned to each object represented by a set of fuzzy features. GNP is one of the evolutionary algorithms and the main object of its evolution in this paper is to discover fuzzy rules from a fuzzy object oriented database. The optimization of the clusters is executed so that the objects with high similarity are put into the same cluster. The results of clustering simulations show that the proposed method can create better clusters comparing to the conventional clustering methods.