Best \(N\) term approximation spaces for tensor product wavelet bases (Q2506201)

The author investigates approximation spaces for best \(N\) term approximation (nonlinear approximation) using tensor product bases of univariate hierarchical wavelet bases. These tensor product bases are anisotropic, i.e., high frequencies in some direction can be tensorized with low frequencies in other direction. Subsets of these bases span the so-called sparse grid space, which can been applied to the numerical solution of elliptic PDEs, to data mining, and to high-dimensional integration.