An approximation method with data selection process (Q730884): Difference between revisions

The authors present a method of curve approximation which derives from an interpolation method of \textit{Ch. P. Bernard} et al. [in: Curve und surface fitting, Saint-Malo 2002. Fifth international conference on curves and surfaces, Saint-Malo, France, June 27 -- July 3, 2002. Modern Methods in Mathematics, 59--64 (2003; Zbl 1037.65014)]. The interpolation method consists in finding an interpolation function as a finite linear combination of certain functions by minimizing a suitable functional and requires a previous selection and discard process of the initial informations. The authors apply the method to the fitting case, by introducing adequate adaptations to the allocation and selection process and minimizing a quadratic functional in a suitable space. They prove that the minimization problem has a unique solution and state convergence results. They also test the method by numerical and graphical examples.