Geometric dual formulation for first-derivative-based univariate cubic L₁ splines

From MaRDI portal
Publication:2481371





Univariate splines can be highly useful for the approximation and smoothing of function and data, where the splines minimizing the \(L_1\)-norm of certain data are especially good when there are abrupt changes in the data. One reason for this is that the \(L_1\)-norm tends to be better when there are a few ``stray data among the input information. In this article, these points are employed to find a good approach by minimizing the \(L_1\)-norm of differences between derivatives of the spline and divided differences of the data. The coefficients of the approximant are computed by a nonsmooth, nonlinear convex optimization problem.



Cites work









This page was built for publication: Geometric dual formulation for first-derivative-based univariate cubic \(L_{1}\) splines

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2481371)