Representation of experimental data by Fourier functions for differentiation

ALEAST-SQUARES approximation routine has been developed to fit Fourier functions to a given set of experimental data in order to obtain the first and second spatial derivatives. It has been observed that a continuous fit can be obtained provided that the input data are given in the form of a smoothly varying periodic function. Furthermore, the number of terms of the expansion must be chosen such that first the fit, and then the two subsequent derivatives are optimized. The derivatives are judged to be optimum when they are free from rapid oscillations. Contents