REMARKS ON ADAPTIVE FOURIER DECOMPOSITION
Abstract
This is a continuation of the study of adaptive Fourier decomposition (AFD).15 Under a mild condition not in terms of smoothness, a convergence rate is provided. We prove that the selection of the parameters corresponding to Fourier series in the average sense is optimal. We also present the transformation matrices between the adaptive rational orthogonal system and the related sequence of the shifted Cauchy kernels and their derivatives.
Supported by University of Macau under research Grant RG056/01-02S/QT/FST.