10 Self-Organized Kernel Fuzzy Clustering Method and Its Evaluation for the Number of Clusters Based on Alignment of Similarities

This paper proposes a self-organized kernel fuzzy clustering method based on the idea of self-organized dissimilarity between two objects, while considering the degree of sharing of common properties of the pair of objects for the obtained fuzzy clusters. Moreover, we propose a method for evaluation of an adaptable number of clusters using an idea of measuring the degree of agreement between the target similarity data and the restored similarity which consists of the result of the self-organized kernel fuzzy clustering model.