A semantic driven evolutive fuzzy clustering algorithm

In this paper it is show that the same semantic constraints used to ensure linguistic interpretation in data driven design of fuzzy systems are also useful in the design of evolutive fuzzy clustering algorithms. Specifically it is show that these constraints generalize the constraints used in popular fuzzy clustering algorithms such as the FCM. Experimental studies illustrate the effectiveness of this approach to clustering. The algorithm attempts to optimize the clusters´ parameters as well as the number of clusters (a dynamically variable length of chromosomes is used).