Gauss-Christoffel quadrature for inverse regression: applications to computer experiments

DOI10.1007/S11222-018-9816-4MaRDI QIDQ2329776zbMATH OpenOpenAlexWikidataFDO

Authors Andrew Glaws, P. Constantine

Publication date 18 October 2019

Published in Statistics and Computing (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1710.01372

zbMATH Keywords

numerical integrationorthogonal polynomialssliced average variance estimationsliced inverse regressionsufficient dimension reduction

Mathematics Subject Classification ID

Computational methods for problems pertaining to statistics (62-08) Factor analysis and principal components; correspondence analysis (62H25) Linear inference, regression (62J99) Numerical integration (65D30)

Abstract: Sufficient dimension reduction (SDR) provides a framework for reducing the predictor space dimension in regression problems. We consider SDR in the context of deterministic functions of several variables such as those arising in computer experiments. In this context, SDR serves as a methodology for uncovering ridge structure in functions, and two primary algorithms for SDR---sliced inverse regression (SIR) and sliced average variance estimation (SAVE)---approximate matrices of integrals using a sliced mapping of the response. We interpret this sliced approach as a Riemann sum approximation of the particular integrals arising in each algorithm. We employ well-known tools from numerical analysis---namely, multivariate tensor product Gauss-Christoffel quadrature and orthogonal polynomials---to produce new algorithms that improve upon the Riemann sum-based numerical integration in SIR and SAVE. We call the new algorithms Lanczos-Stieltjes inverse regression (LSIR) and Lanczos-Stieltjes average variance estimation (LSAVE) due to their connection with Stieltjes' method---and Lanczos' related discretization---for generating a sequence of polynomials that are orthogonal to a given measure. We show that the quadrature-based approach approximates the desired integrals, and we study the behavior of LSIR and LSAVE with three numerical examples. As expected in high order numerical integration, the quadrature-based LSIR and LSAVE exhibit exponential convergence in the integral approximations compared to the first order convergence of the classical SIR and SAVE. The disadvantage of LSIR and LSAVE is that the underlying tensor product quadrature suffers from the curse of dimensionality---that is, the number of quadrature nodes grows exponentially with the input space dimension. Therefore, the proposed approach is most appropriate for deterministic functions with fewer than ten independent inputs.

Recommendations

• Inverse regression for ridge recovery: a data-driven approach for parameter reduction in computer experiments
• Sufficient dimension reduction: methods and applications with R
• Sufficient Dimension Reduction via Inverse Regression
• Kernel dimension reduction in regression
• Dimension reduction in regressions through weighted variance estimation

Cites work

• A Lanczos method for approximating composite functions
• A method for numerical integration on an automatic computer
• A review on dimension reduction
• A taxonomy of global optimization methods based on response surfaces
• Active subspaces. Emerging ideas for dimension reduction in parameter studies
• Approximation theory and approximation practice
• Calculation of Gauss Quadrature Rules
• Comment
• Conditioning as disintegration
• Design and analysis of computer experiments. With comments and a rejoinder by the authors
• Generation and Use of Orthogonal Polynomials for Data-Fitting with a Digital Computer
• Global sensitivity analysis: The primer
• scientific article; zbMATH DE number 1209388 (Why is no real title available?)
• scientific article; zbMATH DE number 1220060 (Why is no real title available?)
• scientific article; zbMATH DE number 1246413 (Why is no real title available?)
• scientific article; zbMATH DE number 2107939 (Why is no real title available?)
• scientific article; zbMATH DE number 961607 (Why is no real title available?)
• Inverse regression for ridge recovery: a data-driven approach for parameter reduction in computer experiments
• Inverse regression-based uncertainty quantification algorithms for high-dimensional models: theory and practice
• Krylov subspace methods. Principles and analysis.
• Likelihood-based sufficient dimension reduction
• Matrices, moments and quadrature with applications
• On Directional Regression for Dimension Reduction
• On Principal Hessian Directions for Data Visualization and Dimension Reduction: Another Application of Stein's Lemma
• Principal component analysis.
• Regression analysis under link violation
• Response surface methodology. Process and product optimization using designed experiments
• Ridge functions
• Save: a method for dimension reduction and graphics in regression
• Sliced Inverse Regression for Dimension Reduction
• Sparse pseudospectral approximation method
• Sufficient dimension reduction and prediction in regression
• Sufficient Dimension Reduction via Inverse Regression
• The design and analysis of computer experiments.
• The interplay between classical analysis and (numerical) linear algebra -- a tribute to Gene H. Golub
• The Lanczos and Conjugate Gradient Algorithms
• The numerically stable reconstruction of a Jacobi matrix from spectral data

Cited in

(3)

• Embedded ridge approximations
• Inverse regression for ridge recovery: a data-driven approach for parameter reduction in computer experiments
• Explainable binary classification of separable shape ensembles

Describes a project that uses

Uses Software

• Virtual library

This page was built for publication: Gauss-Christoffel quadrature for inverse regression: applications to computer experiments

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2329776)