Results 1  10
of
43
Kernel Techniques: From Machine Learning to Meshless Methods
, 2006
"... Kernels are valuable tools in various fields of Numerical Analysis, including approximation, interpolation, meshless methods for solving partial differential equations, neural networks, and Machine Learning. This contribution explains why and how kernels are applied in these disciplines. It uncovers ..."
Abstract

Cited by 30 (8 self)
 Add to MetaCart
Kernels are valuable tools in various fields of Numerical Analysis, including approximation, interpolation, meshless methods for solving partial differential equations, neural networks, and Machine Learning. This contribution explains why and how kernels are applied in these disciplines. It uncovers the links between them, as far as they are related to kernel techniques. It addresses nonexpert readers and focuses on practical guidelines for using kernels in applications.
Solving Differential Equations with Radial Basis Functions: Multilevel Methods and Smoothing
 Advances in Comp. Math
"... . Some of the meshless radial basis function methods used for the numerical solution of partial differential equations are reviewed. In particular, the differences between globally and locally supported methods are discussed, and for locally supported methods the important role of smoothing within a ..."
Abstract

Cited by 29 (7 self)
 Add to MetaCart
. Some of the meshless radial basis function methods used for the numerical solution of partial differential equations are reviewed. In particular, the differences between globally and locally supported methods are discussed, and for locally supported methods the important role of smoothing within a multilevel framework is demonstrated. A possible connection between multigrid finite elements and multilevel radial basis function methods with smoothing is explored. Various numerical examples are also provided throughout the paper. 1. Introduction During the past few years the idea of using socalled meshless methods for the numerical solution of partial differential equations (PDEs) has received much attention throughout the scientific community. As a few representative examples we mention Belytschko and coworker's results [3] using the socalled elementfree Galerkin method, Duarte and Oden's work [11] using hp clouds, Babuska and Melenk 's work [1] on the partition of unity method, ...
Error Bounds for Solving Pseudodifferential Equations on Spheres by Collocation with Zonal Kernels
 J. Approx. Theory
, 1999
"... . The problem of solving pseudodifferential equations on spheres by collocation with zonal kernels is considered and bounds for the approximation error are established. The bounds are given in terms of the maximum separation distance of the collocation points, the order of the pseudodifferential ..."
Abstract

Cited by 16 (1 self)
 Add to MetaCart
. The problem of solving pseudodifferential equations on spheres by collocation with zonal kernels is considered and bounds for the approximation error are established. The bounds are given in terms of the maximum separation distance of the collocation points, the order of the pseudodifferential operator, and the smoothness of the employed zonal kernel. A byproduct of the results is an improvement on the previously known convergence order estimates for Lagrange interpolation. 2000 Mathematics Subject Classification. Primary 41A25, 65N15, 65N35; Secondary 41A05, 41A30, 41A63, 58J40, 65D05. Key words and phrases. Pseudodifferential equation, collocation, zonal kernel, interpolation, approximation order, sphere, positive definite function, radial basis function. x1. Introduction Data fitting and solving differential and integral equations on the sphere are areas of growing interest with applications to physical geodesy, potential theory, oceanography, and meteorology [6,10]. As...
Relationship between boundary integral equation and radial basis function
 in The 52th Symposium of JSCME on BEM
, 2000
"... This paper aims to survey our recent work relating to the radial basis function (RBF) from some new views of points. In the first part, we established the RBF on numerical integration analysis based on an intrinsic relationship between the Green's boundary integral representation and RBF. It is ..."
Abstract

Cited by 10 (10 self)
 Add to MetaCart
This paper aims to survey our recent work relating to the radial basis function (RBF) from some new views of points. In the first part, we established the RBF on numerical integration analysis based on an intrinsic relationship between the Green's boundary integral representation and RBF. It is found that the kernel function of integral equation is important to create efficient RBF. The fundamental solution RBF (FSRBF) was presented as a novel strategy constructing operatordependent RBF. We proposed a conjecture formula featuring the dimension affect on error bound to show the independentdimension merit of the RBF techniques. We also discussed wavelet RBF, localized RBF schemes, and the influence of node placement on the RBF solution accuracy. The centrosymmetric matrix structure of the RBF interpolation matrix under symmetric node placing is proved. The second part of this paper is concerned with the boundary knot method (BKM), a new boundaryonly, meshless, spectral convergent, integrationfree RBF collocation technique. The BKM was tested to the Helmholtz, Laplace, linear and nonlinear convectiondiffusion problems. In particular, we introduced the response knotdependent nonsingular general solution to calculate varyingparameter and nonlinear steady convectiondiffusion problems very efficiently. By comparing with the multiple dual reciprocity method, we discussed the completeness issue of the BKM. Finally, the nonsingular solutions for some known differential operators were given in appendix. Also we expanded the RBF concepts by introducing timespace RBF for transient problems. Key words: radial basis function, integral equation, kernel function, edge effects, FSRBF, error bound, node placement, timespace
Multivariate Splines for Data Fitting and Approximation
"... Abstract. Methods for scattered data fitting using multivariate splines will be surveyed in this paper. Existence, uniqueness, and computational algorithms for these methods, as well as their approximation properties will be discussed. Some applications of multivariate splines for data fitting will ..."
Abstract

Cited by 9 (7 self)
 Add to MetaCart
Abstract. Methods for scattered data fitting using multivariate splines will be surveyed in this paper. Existence, uniqueness, and computational algorithms for these methods, as well as their approximation properties will be discussed. Some applications of multivariate splines for data fitting will be briefly explained. Some new research initiatives of scattered data fitting will be outlined. Given a set of scattered data, e.g., {(xi, yi, zi), i = 1, · · · , N}, we need to find a smooth function or surface S such that S(xi, yi) = zi, i = 1, · · · , N, if zi are very accurate measurements or
Polynomial operators and local approximation of solutions of pseudodifferential equations on the sphere, Submitted for publication
"... We study the solutions of an equation of the form Lu = f, where L is a pseudodifferential operator defined for functions on the unit sphere embedded in a Euclidean space, f is a given function, and u is the desired solution. We give conditions under which the solution exists, and deduce local smoo ..."
Abstract

Cited by 9 (8 self)
 Add to MetaCart
(Show Context)
We study the solutions of an equation of the form Lu = f, where L is a pseudodifferential operator defined for functions on the unit sphere embedded in a Euclidean space, f is a given function, and u is the desired solution. We give conditions under which the solution exists, and deduce local smoothness properties of u given corresponding local smoothness properties of f, measured by local Besov spaces. We study the global and local approximation properties of the spectral solutions, describe a method to obtain approximate solutions using values of f at points on the sphere and polynomial operators, and describe the global and local rates of approximation provided by our polynomial operators. 1
Spherical Splines for Data Interpolation and Fitting
 SIAM J. Scientific Computing
, 2005
"... Abstract. We study minimal energy interpolation, discrete and penalized least squares approximation problems on the unit sphere using nonhomogeneous spherical splines. Several numerical experiments are conducted to compare approximating properties of homogeneous and nonhomogeneous splines. Our numer ..."
Abstract

Cited by 9 (4 self)
 Add to MetaCart
(Show Context)
Abstract. We study minimal energy interpolation, discrete and penalized least squares approximation problems on the unit sphere using nonhomogeneous spherical splines. Several numerical experiments are conducted to compare approximating properties of homogeneous and nonhomogeneous splines. Our numerical experiments show that nonhomogeneous splines have certain advantages over homogeneous splines. Key words. spherical splines, data fitting AMS subject classifications. 65D05, 65D07, 65D17
M.J.: Stable evaluation of Gaussian RBF interpolants
 In preparation
"... Abstract. We provide a new way to compute and evaluate Gaussian radial basis function interpolants in a stable way also for small values of the shape parameter, i.e., for “flat ” kernels. This work is motivated by the fundamental ideas proposed earlier by Bengt Fornberg and his coworkers. However, ..."
Abstract

Cited by 5 (3 self)
 Add to MetaCart
(Show Context)
Abstract. We provide a new way to compute and evaluate Gaussian radial basis function interpolants in a stable way also for small values of the shape parameter, i.e., for “flat ” kernels. This work is motivated by the fundamental ideas proposed earlier by Bengt Fornberg and his coworkers. However, following Mercer’s theorem, an L2(Rd,ρ)orthonormal expansion of the Gaussian kernel allows us to come up with an algorithm that is simpler than the one proposed by Fornberg, Larsson and Flyer and that is applicable in arbitrary space dimensions d. In addition to obtaining an accurate approximation of the RBF interpolant (using many terms in the series expansion of the kernel) we also propose and investigate a highly accurate leastsquares approximation based on early truncation of the kernel expansion. Key words. Radial basis functions, Gaussian kernel, stable evaluation, QR decomposition.
APPROXIMATION ON THE SPHERE USING RADIAL BASIS FUNCTIONS PLUS POLYNOMIALS
"... Abstract. In this paper we analyse a hybrid approximation of functions on the sphere S 2 ⊂ R 3 by radial basis functions combined with polynomials, with the radial basis functions assumed to be generated by a (strictly) positive definite kernel. The approximation is determined by interpolation at sc ..."
Abstract

Cited by 5 (1 self)
 Add to MetaCart
(Show Context)
Abstract. In this paper we analyse a hybrid approximation of functions on the sphere S 2 ⊂ R 3 by radial basis functions combined with polynomials, with the radial basis functions assumed to be generated by a (strictly) positive definite kernel. The approximation is determined by interpolation at scattered data points, supplemented by side conditions on the coefficients to ensure a square linear system. The analysis is first carried out in the native space associated with the kernel (with no explicit polynomial component, and no side conditions). A more refined error estimate is obtained for functions in a still smaller space. Numerical calculations support the utility of this hybrid approximation. Key words. Scattered data, radial basis functions, spherical harmonics, error estimate. AMS subject classifications. 41A30, 65D30 1. Introduction. In