谱聚类能够识别任意形状的样本空间且收敛于全局最优解，其基本思想是利用样本数据相似矩阵的进行特征分解后得到的特征向量进行聚类，程序进行了几种不同聚类算法的比较，包括Q矩阵聚类，kmeans聚类，第一特征分量聚类，第二广义特征分量聚类，公用数据生成和近邻矩阵生成(Spectral clustering can distinguish arbitrary sample space and converge to the global optimal solution, the basic idea is similar to matrix of eigenvectors obtained after decomposition of clustering using the sample data, procedures for the comparison of several different clustering algorithms, including Q matrix clustering, kmeans clustering, the first feature clustering, second generalized characteristic component clustering, public data generation and neighbor matrix generation)