Approximation of eigenfunctions in kernel-based spaces (Q314464): Difference between revisions

The authors study the approximation of eigenfunctions in kernel-based spaces. They use the approximation error that is bounded in terms of generalized power functions and use determination of error-optimal \(n\)-dimensional subspaces. The authors prove that eigenspaces minimize the \(L_{2}(\Omega)\) norm of the power function. Several examples are given to illustrate the numerical technique via a greedy point selection strategy. In addition they show that the algorithm provided allows to approximate the eigenvalues for Sobelev spaces in a way that recovers the true decay rates.