Learning from examples with spatial-adaptive wavelet-based reproducing kernels

This paper formulates the problem of learning from examples as a scattered data interpolation problem, and develops a new method that computes interpolants that minimize a wavelet-based reproducing kernel Hilbert space (RKHS) norm subject to interpolatory constraints. In contrast to radial basis function kernels, these kernels are not translation invariant. Some computational geometry methods are used to construct spatial-adaptive kernels based on local distribution density of unevenly distributed data examples