Clustering in image space in support vector machine

The kernel-based clustering has attracted great attention with the development of support vector machine. One can perform a clustering approach in an image space after mapping the data in an original space to the image space, but it is difficult to capture the optimal parameters for finding real clusters. In this paper, we present a kernel-based clustering approach in light of a relational fuzzy clustering procedure. This approach offers a better solution to the kernel-based clustering compared with conventional approaches. Experiments are presented to demonstrate the effectiveness of our proposed method.