Kernel-Based Fuzzy Competitive Learning Clustering
Kiyotaka Mizutani, Satoshi Miyamoto · 2005
Clustering by competitive learning has been often studied as one of unsupervised classification methods, and some clustering algorithms using a kernel trick employed in nonlinear transformation into a high-dimensional feature space in the support vector machines have been studied to obtain nonlinear cluster boundaries. This paper aims at proposing an algorithm of fuzzy competitive learning clustering using kernel function, and derivation of a fuzzy classification function. Numerical examples are shown and effect of the kernel-based method is discussed