基于全局K-means的谱聚类算法

Spectral clustering based on global K-means

谱聚类算法是近年来研究得比较多的一种聚类算法。但谱聚类是对初始化敏感的,针对这种缺陷,提出一种基于全局K-means的谱聚类算法(GKSC),引入对初值不敏感的全局K-means算法来改善。通过仿真实验表明:GKSC与传统谱聚类算法相比更能得到稳定的聚类结果和更高的聚类精确度。

Spectral clustering is a newly emerged effective and widely used clustering method.With the essence of initialization sensitivity in spectral clustering,the global K-means clustering algorithm was introduced.Then a spectral clustering algorithm based on global K-means was proposed.Compared with the traditional spectral algorithm,some experiments show that the proposed algorithm is not only effective and feasible but also good at getting stable clustering results and of high clustering precision.