The kernel methodKernel method is a versatile technique for extending linear model processing algorithms to nonlinear models. It became well-known in conjunction with the popularization of the support vector machine, a method of pattern recognition. Its most significant feature is that by introducing a function known as the kernel function, it can compute the inner product of vectors nonlinearly mapped to high-dimensional spaces using operations in the low-dimensional original space. As a result, it became possible to naturally extend linear processing algorithms for multivariate data to nonlinear ones. Currently, the kernel method is becoming one of the fundamental techniques not only in pattern recognition and machine learning but also in computer vision, natural language processing, data mining, and bioinformatics, among other broad application fields.