说明：  将数据集转换为拉普拉斯矩阵，然后利用基于图论的谱聚类进行聚类。拉普拉斯矩阵采用高斯核函数，全连接方法计算。谱聚类擅长处理高维数据或非凸数据集。
(The data set is transformed into Laplacian matrix, and then clustered by spectral clustering based on graph theory. The Laplacian matrix is calculated by using the Gauss kernel function and the full connection method. Spectral clustering is good at dealing with high-dimensional or non-convex data sets.)