Wavelet denoising for differential operator

The problem of model selection has attracted the attention of both applied and theoretical statistics for as long as one can remember. But some interesting scientific applications involve indirect noisy measurements. This paper deals with a realizable adaptive estimator for the treatment of a derivative from noisy data. It uses the basic ingredients of a multiresolution construction which is well known as non-standard forms derived by Beylkin etc. The methodology for the derivative denoising was extended in [7, 8]. It can be implemented fast and generalized to the whole family of integral operators with integral kernels as well. In this paper, we focus our attention on the numerical test and implementation technical details of the proposed approach.