6.Comparatively, for SPCA using Gauss KernelFunction and Laplace KernelFunction, it is required to normalize original data according to the category, and the constructed evaluation functions are better than the ones constructed via using PCA.
10.The experiment results show that,when employing the same kernelfunction,the self-adaptability and forecasting accuracy of FFLSSVM is superior to SVM and LSSVM.