Review on spectral methods for clustering
Jingmao Zhang, Shen YanXia · 2015
Spectral clustering(SC) is a clustering technology based on graph theory. It becomes one of the most hot topics on clustering because that it can get global optimal solution without any assumptions on data's structure. In this paper, the basic graph theories including some typical graph cut methods for SC are described, then, classic SC algorithms are introduced. Several problems and research topics on SC are also predicted at the end of this paper.