KERNEL CLUSTERING ALGORITHM
Z Li, Da Zhang, Jinfeng Cheng · Chinese Journal of Computers · 2002
A new clustering algorithm is proposed for cluster analysis in this paper. In general, the reliability of the traditional clustering algorithms strictly depends on the feature difference of data. If the feature differences are large, it is easy to implement clustering. But if the feature differences are small and even cross in the origin space, it is difficult for traditional algorithms to clustering correctly. We adopt the traditional clustering methods and the kernel technique to construct our kernel clustering algorithm. By using Mercer kernel functions, we can map the data in the original space to a high dimensional feature space in which we can perform clustering efficiently. The features of kernel clustering algorithm are fast in convergence speed and accurate in clustering, compared with classical clustering algorithms. The results of simulation experiments show the feasibility and effectiveness of the kernel clustering algorithm.