SPCA arithmetic has smaller number of spectral principal components and greater variance contribution than PCA by choosing proper kernel functions and parameters.

• 当选取合适的核函数和参数时，谱主成分的个数比主成分的个数要小且累积方差贡献率要大。