Variable Transformations in combination with Wavelets and ANOVA for high-dimensional approximation

Daniel Potts; Laura Weidensager

Home/Archive/arXiv

arXiv · 2207.12826

Variable Transformations in combination with Wavelets and ANOVA for high-dimensional approximation

Daniel Potts·Laura Weidensager

View original source ↗Read source document ↗Export metadata ↓

Abstract

We use hyperbolic wavelet regression for the fast reconstruction of high-dimensional functions having only low dimensional variable interactions. Compactly supported periodic Chui-Wang wavelets are used for the tensorized hyperbolic wavelet basis on the torus. With a variable transformation we are able to transform the approximation rates and fast algorithms from the torus to other domains. We perform and analyze scattered-data approximation for smooth but arbitrary density functions by using a least squares method. The corresponding system matrix is sparse due to the compact support of the wavelets, which leads to a significant acceleration of the matrix vector multiplication. For non-periodic functions we propose a new extension method. A proper choice of the extension parameter together with the piece-wise polynomial Chui-Wang wavelets extends the functions appropriately. In every case we are able to bound the approximation error with high probability. Additionally, if the function has low effective dimension (i.e. only interactions of few variables), we qualitatively determine the variable interactions and omit ANOVA terms with low variance in a second step in order to decrease the approximation error. This allows us to suggest an adapted model for the approximation. Numerical results show the efficiency of the proposed method.

Explore related subjects

math.NA cs.NA

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Daniel Potts, Laura Weidensager. 2024-05-29. Variable Transformations in combination with Wavelets and ANOVA for high-dimensional approximation. https://arxiv.org/abs/2207.12826

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Puzzle Game: Prediction and Classification of Wordle Solution Words

Wordle result prediction involves both daily participation and the difficulty of the solution word. We model these quantities with ARIMA forecasts of report counts, XGBoost regressors using word frequency, letter entropy, and repetition, and a decision tree trained on K-means difficulty tiers. Experiments use 355 daily outcome distributions, with a 107-word chronological holdout and four expanding-window tests. XGBoost reduces holdout mean absolute error from 4.08 to 3.84 percentage points and improves over the training mean in every temporal window. Ridge and random forests remain competitive. XGBoost systematically underestimates failure rates on difficult words. The lexical decision tree achieves 54.2% tier accuracy, compared with 40.2% when predicted histograms are assigned to the same training centroids. Bootstrap analysis shows that cluster stability does not imply that tiers are predictable from lexical attributes. For report counts, comparisons at matched horizons reveal substantial overdispersion in full-history ARIMA intervals. These results characterize the predictive value and limitations of a compact lexical model of reported Wordle performance.

math.NA↗

arXiv2026

Nodal Coarsening and Approximate Ideal Interpolation for H(curl) in Algebraic Multigrid

We propose an approximate interpolation construction and a practical coarsening algorithm for algebraic multigrid (AMG) method when applied to H(curl) problems. Building on the generalized AMG framework, we introduce an interior/exterior splitting that yields both a refinement-based and a fully algebraic construction of the interpolation. The refinement-based approach follows a geometric hierarchy, while the purely algebraic interpolation is constructed through a coarsening process that first coarsens a nodal dual problem and then builds coarse and fine variables using a matching algorithm. Under a uniform fine-space smoothing assumption, we prove a jump-independent two-grid convergence bound for canonical finite-element spaces. We also establish an exact commuting relation for the unshifted construction. Two-dimensional numerical experiments demonstrate sparse multilevel hierarchies and robust convergence under strong coefficient jumps in the curl--curl term. Three-dimensional experiments provide a two-level proof of concept, although the interpolation operators remain dense. Developing a practical sparse three-dimensional multilevel method is left to future work.

math.NA↗

arXiv2026

Local discontinuous Galerkin FEM for convex minimization

The heart of the a priori and a posteriori error control in convex minimization problems is the sharp control of the differences of discrete and exact minimal energy. Conforming finite element discretizations for p-Laplace type minimization problems provide upper bounds of the energy difference with optimal convergence rates. Even for smooth solutions, known convergence rates for higher-order non-conforming finite element discretizations for the same problem class with $2 < p < \infty$, however, are exclusively suboptimal. Thus the popular a posteriori error control within the two-energy principle, that generalize hyper-circle identities, appears unbalanced. The innovative point of departure in a refined analysis of two discontinuous Galerkin (dG) schemes exploits duality relations between a discrete primal and a semi-discrete dual problem. The infinite-dimensional dual problem leads to a tiny duality gap that even vanishes for polynomial low-order terms. For a class of degenerated convex minimization problems with two-sided $p$ growth, the novel duality provides improved a priori convergence rates for the error in the minimal energies. This closes the misfit of convergence rates for the conforming and nonconforming schemes at least for the local discontinuous Galerkin schemes at hand. The motivating two-energy principle and some post-processing for a Raviart-Thomas dual variable provides an a posteriori error control, that also may drive adaptive mesh-refining. Computational benchmarks provide numerical evidence for improved convergence rates of the adaptive beyond uniform mesh-refining.

math.NA↗