A self-adaptive spectral clustering algorithm
Xiaoyan Cai, Guanzhong Dai, Libin Yang, Guoqing Zhang · 2008
Most existing algorithms on spectral clustering are not able to determine the number of clusters. In this paper, we prove theoretically that the eigenvectors of the affinity matrix can be used directly to cluster the data points. And we suggest exploiting the structure of the eigenvectors to infer automatically the number of clusters. As a result, a self-adaptive spectral clustering algorithm based on affinity matrix is proposed. The experimental results on the UCI data sets show that the algorithm is more effective than previous algorithms.