A consideration on vector quantization clusterings with kernel functions
Tomiyuki Shiiba, Noritaka Shigei, Hiromi Miyajima · ITC-CSCC :International Technical Conference on Circuits Systems, Computers and Communications · 2009
Kernel methods are ones that, by replacing the inner product with positive definite function, implicitly perform a nonlinear mapping of the input data into a high-dimensional feature space. So far, a large number of vector quantization methods such as Learning vector quantization (LVQ) and Neural gas (NG) with kernel functions have been proposed. But, the performance of them has not always clarified yet. In this paper, we perform some simulations of kernel clusterings with small and large number of weight vectors. As a result, it is show that the performance of LVQ and NG with kernel function are the same as one of them with a large number of weight vectors.