Polynomial approximation of noisy functions
As known, the generic algorithm for the least-squares problem requires \(O(N n^2)\) operations, where \(N+1\) is the number of sample points and \(n\) is the degree of the polynomial approximant. The authors propose an \(O(N \log N)\) method for polynomial approximation of a univariate noisy function. The method is based on truncating the Chebyshev interpolant at an appropriate degree and corresponds to solving a weighted least-squares problem. Blending numerical analysis and statistics, the convergence of the method is also examined. The convergence is spectral until the error reaches \(O(\sigma \sqrt{n/N})\), where \(\sigma\) is the noise level, after which the error continues to decrease at the Monte-Carlo \(O(1/\sqrt{N})\) rate. The Mallows' \(C_p\) statistical tool is used to determine the polynomial degree without prior knowledge of the noise level \(\sigma\).\N\NThe method, called NoisyChebtrunc combines the computational efficiency with \(O(N \log N)\) operations, the stability inherent in Chebyshev interpolation, leading to \(L_{\infty}\) convergence results with high probability and the Monte-Carlo style noise reduction as more samples are taken.\N\NThe efficiency of NoisyChebtrunc is illustrated with numerical experiments.
- A consistency property of the AIC for multivariate linear models when the dimension and the sample size are large
- Accurate derivative estimation from noisy data: a state-space approach
- All of Nonparametric Statistics
- An Extension of Chebfun to Two Dimensions
- An introduction to statistical learning. With applications in R
- An unbiased \(C_p\) criterion for multivariate ridge regression
- APPROXIMATING SMOOTH, MULTIVARIATE FUNCTIONS ON IRREGULAR DOMAINS
- Approximation in the extended functional tensor train format
- Asymptotic behavior of M-estimators for the linear model
- Asymptotic behavior of M-estimators of p regression parameters when p^ 2/n is large. I. Consistency
- Chopping a Chebyshev Series
- Consistency of high-dimensional AIC-type and C_p-type criteria in multivariate linear regression
- Convergence estimates in probability and in expectation for discrete least squares with noisy evaluations at random points
- Data-driven algorithms for signal processing with trigonometric rational functions
- Error bounds for derivative estimates based on spline smoothing of exact or noisy data
- Frames and numerical approximation
- Gaussian model selection
- Géza Freud, orthogonal polynomials and Christoffel functions. A case study
- High-dimensional probability. An introduction with applications in data science
- High-dimensional statistics. A non-asymptotic viewpoint
- scientific article; zbMATH DE number 1156422 (Why is no real title available?)
- scientific article; zbMATH DE number 6438182 (Why is no real title available?)
- Information criteria and statistical modeling.
- Introduction to nonparametric estimation
- LSQR: An Algorithm for Sparse Linear Equations and Sparse Least Squares
- Numerical Methods in Scientific Computing, Volume I
- On the stability and accuracy of least squares approximations
- Optimal weighted least-squares methods
- Pattern recognition and machine learning.
- Some Comments on C P
- Some new asymptotic theory for least squares series: pointwise and uniform results
- Sparse grids
- Spectral Methods in MATLAB
- The AAA algorithm for rational approximation
- The Discrete Cosine Transform
This page was built for publication: Polynomial approximation of noisy functions
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6926533)
