Fast clustering with application to fuzzy rule generation

This paper presents a multistage random sampling fuzzy c-means based clustering algorithm, which significantly reduces the computation time required to partition a data set into c classes. A series of subsets of the full data set are used for classification in order to provide an approximation to the final cluster centers. The quality of the final partitions is equivalent to that of fuzzy c-means. The speed-up is normally a factor of 2-3 times, which is especially significant for high dimensional spaces and large data sets. Clustering has been used to generate fuzzy rules for control. In this paper, we show that the multistage random sampling fuzzy c-means based clustering algorithm can be effectively used to create fuzzy rules in the domain of magnetic resonance images where over 60,000 patterns and 3 features or attributes are common