Abstract:This paper studies function approximation in Gaussian Sobolev spaces over the real line and measures the error in a Gaussian-weighted $L^p$-norm. We construct two linear approximation algorithms using $n$ function evaluations that achieve the optimal or almost optimal rate of worst-case convergence in a Gaussian Sobolev space of order $\alpha$. The first algorithm is based on scaled trigonometric interpolation and achieves the optimal rate $n^{-\alpha}$ up to a logarithmic factor. This algorithm can be constructed in almost-linear time with the fast Fourier transform. The second algorithm is more complicated, being based on spline smoothing, but attains the optimal rate $n^{-\alpha}$.

What problem does this paper attempt to address?

### What problems does this paper attempt to solve? This paper mainly studies the problem of function approximation in the Gaussian Sobolev space and uses the Gaussian - weighted $ L^p $-norm to measure the error. Specifically, the authors construct two linear approximation algorithms. These algorithms recover the function $ f: \mathbb{R} \to \mathbb{R} $ by evaluating $ n $ points of the function and achieve the optimal or near - optimal worst - case convergence rate in the Gaussian Sobolev space $ W^{\alpha, q}_\rho $. #### Main problems: 1. **Function approximation problem**: How to construct a linear approximation algorithm $ A_n(f) $ in the Gaussian Sobolev space $ W^{\alpha, q}_\rho $ so that it can recover a given function $ f $ as accurately as possible through a finite number of function evaluations (i.e., $ n $ times). 2. **Worst - case convergence rate**: Ensure that the constructed algorithm can reach the optimal or near - optimal worst - case convergence rate $ n^{-\alpha} $, where $ \alpha $ is the order of the Sobolev space. #### Specific goals: - **The first algorithm**: Based on the scaled triangular interpolation, this algorithm can achieve the optimal convergence rate $ n^{-\alpha} $ (up to a logarithmic factor) and can be constructed in almost linear time through the fast Fourier transform (FFT). - **The second algorithm**: Based on spline smoothing, this algorithm is more complex but can reach the strictly optimal convergence rate $ n^{-\alpha} $. #### Research background: - In the high - dimensional case, function approximation is an important task. Recent research has shown that in the one - dimensional Gaussian Sobolev space $ W^{\alpha, q}_\rho $, the best possible sampling recovery convergence rate is $ n^{-\alpha} $, but no clear construction method has been given for specific algorithm implementation. - This paper fills this gap and provides a basis for subsequent research in the high - dimensional case. By solving these problems, the paper not only provides new insights for theoretical research but also provides practical algorithm tools for efficient function approximation in practical applications.

Construction of Optimal Algorithms for Function Approximation in Gaussian Sobolev Spaces

Approximation in Hilbert spaces of the Gaussian and related analytic kernels

Optimal numerical integration and approximation of functions on $\mathbb{R}^d$ equipped with Gaussian measure

On Nonlinear Simultaneous Approximation Problems

Optimal approximation of elliptic problems by linear and nonlinear mappings II

Optimal Piecewise Linear Function Approximation for GPU-based Applications

Multi- and Infinite-variate Integration and $L^2$-Approximation on Hilbert Spaces with Gaussian Kernels

Exponential Convergence of an Approximation Problem for Infinitely Differentiable Multivariate Functions

Learning Lipschitz Operators with respect to Gaussian Measures with Near-Optimal Sample Complexity

Best One-Sided Approximation and Optimal Hermite-Fejer Interpolation on an Infinite Interval

On Interpolation Approximation: Convergence Rates for Polynomial Interpolation for Functions of Limited Regularity.

Optimal quadrature errors and sampling numbers for Sobolev spaces with logarithmic perturbation on spheres

Line-search based optimization using function approximations with tunable accuracy

Approximation Algorithms for the Gromov Hyperbolicity of Discrete Metric Spaces.

Infinite-Variate $L^2$-Approximation with Nested Subspace Sampling

Optimal pointwise error estimates for piecewise functions expanded with Laguerre polynomials

Approximation Algorithms for Homogeneous Polynomial Optimization with Quadratic Constraints

Constructive Approximation of High-Dimensional Functions with Small Efficient Dimension with Applications in Uncertainty Quantification

Randomized least-squares with minimal oversampling and interpolation in general spaces

Approximation Algorithms for Discrete Polynomial Optimization

On optimal convergence rates of Laguerre polynomial expansions for piecewise functions