Representation of sparse Legendre expansions
The sparse Legendre representation is of the form \(f(t)=\sum_{j=1}^M c_j P_{e_j}(t)\) where \(P_{e_j}\) represents a Legendre polynomial of degree \(e_j\). In this paper, an algorithm is given to compute the degrees \(0\leq e_1<e_2<\cdots<e_M=N\) and the coefficients \(\{c_j\}_{j=1}^M\) where \(M\ll N\), given the values \(f(1),f'(1),\ldots,f^{(2M-1)}(1)\). The \(e_j\) are computed from the values \(z_j=e_j(e_j+1)/2\). The latter appear naturally in the process of taking derivatives. The \(z_j\) are found as the zeros of a polynomial whose coefficients can be computed by solving a linear Hankel system. Once the degrees \(e_j\) are known, the coefficients \(c_j\) can also be computed by solving a linear system.
- Reconstruction of sparse Legendre and Gegenbauer expansions
- Rapidly computing sparse Legendre expansions via sparse Fourier transforms
- A fast and simple algorithm for the computation of Legendre coefficients
- On the convergence rates of Legendre approximation
- Sparse Legendre expansions via _1-minimization
- The generalized operator based Prony method
- Sparse polynomial interpolation in Chebyshev bases
- Reconstruction of sparse Legendre and Gegenbauer expansions
- Sparse Legendre expansions via _1-minimization
- Sparse sums with bases of Chebyshev polynomials of the third and fourth kind
- Sparse polynomial interpolation with Bernstein polynomials
- Rapidly computing sparse Legendre expansions via sparse Fourier transforms
- Nonlinear approximation in bounded orthonormal product bases
- Multiscale matrix pencils for separable reconstruction problems
This page was built for publication: Representation of sparse Legendre expansions
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1930163)