Dimension-dependent error estimates for sampling recovery on Smolyak grids based on B-spline quasi-interpolation (Q2286231): Difference between revisions

Quasi-interpolation and, as an example, quasi-interpolation using (univariate) B-splines is an important and useful tool in approximation theory. In this article, B-spline quasi-interpolation sampling representation is used for dimension-dependent estimates of the approximation error of linear sampling algorithms. These are -- in general -- upper bounds, and they are applied to periodic data. The approximands are from Hölder-Zygmund classes of positive mixed smoothness. There are error estimates for all these approximands, and there are upper and lower bounds if the smoothness is at most two. Examples include functions on unit balls with homogeneous side-conditions and functions with a number of active variables at most as many as the dimension of the space. For these estimates, Smolyak grids are used.