### References & Citations

# Mathematics > Numerical Analysis

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

(Submitted on 30 Jan 2023)

Abstract: We consider $L^2$-approximation on weighted reproducing kernel Hilbert spaces of functions depending on infinitely many variables. We focus on unrestricted linear information, admitting evaluations of arbitrary continuous linear functionals. We distinguish between ANOVA and non-ANOVA spaces, where, by ANOVA spaces, we refer to function spaces whose norms are induced by an underlying ANOVA function decomposition. In ANOVA spaces, we prove that there is an optimal algorithm to solve the approximation problem using linear information. This way, we can determine the exact polynomial convergence rate of $n$-th minimal worst-case errors. For non-ANOVA spaces, we also establish upper and lower error bounds. Even though the bounds do not match in this case, they reveal that for weights with a moderate decay behavior, the convergence rate of $n$-th minimal errors is strictly higher in ANOVA than in non-ANOVA spaces.

Link back to: arXiv, form interface, contact.