Support vector regression with ANOVA decomposition kernels

摘要: Support Vector Machines using ANOVA Decomposition Kernels (SVAD) [Vapng] are a way of imposing structure on multi-dimensional kernels which generated as the tensor product one-dimensional kernels. This gives more accurate control over capacity learning machine (VCdimension). SVAD uses ideas from decomposition methods and extends them to generate directly implement these ideas. is used with spline results show that performs better than respective non kernel. The Boston housing data set UCI has been tested Bagging [Bre94] before [DBK97] compared method.