Inf-convolution and radial basis functions
Let \(\{x_1, x_2, \dots, x_N \}\) be a set of \(N\) distinct points in \({\mathbb R}^d\), let \(z_1, z_2, \ldots ,z_N\) be real numbers, and let \[ \sigma(\cdot)= \sum_{i=1}^N \alpha_i\phi(\|\cdot-x_i\|) +\sum_{j=1}^M\beta_j v_j(\cdot), \] where \(\phi:{\mathbb R}^+ \to {\mathbb R}\) and \(v_j:{\mathbb R}^d \to {\mathbb R}\) are continuous functions. The coefficients \({a_i}\) and \({\beta_j}\) are to be determined from the interpolation conditions \(\sigma(x_k)=z_k\) \((k=1, \ldots, N)\) and the orthogonality conditions \(\sum_{k=1}^N\alpha_kv_j(x_k)=0\) \((j=1, \ldots , M)\). There are functions \(\phi\) of various types for which the above interpolation problem is solvable for every given \(\{x_i\}\) and \(\{z_i\}\). In this paper a method is proposed for constructing compactly supported \(\phi\) having this property. The method is based on the concept of the so-called inf-convolution originated in the convex optimization theory. For the entire collection see [Zbl 0919.00054].
- scientific article; zbMATH DE number 4178625
- Fundamental Solutions of PDEs as Radial Basis Functions in Multivariate Interpolation
- Convergence properties of radial basis functions
- A new class of radial basis functions with compact support
- Multivariate cardinal interpolation with radial-basis functions
- Fast convolution with radial kernels at nonequispaced knots
- Generation of point sets by convex optimization for interpolation in reproducing kernel Hilbert spaces
- Convexity and solvability for compactly supported radial basis functions with different shapes
- Mollification formulas and implicit smoothing
- scientific article; zbMATH DE number 733432 (Why is no real title available?)
- A new class of radial basis functions with compact support
This page was built for publication: Inf-convolution and radial basis functions
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4699395)
