Spectrahedral Regression

DOI10.1137/21M1455899MaRDI QIDQ6155881zbMATH OpenOpenAlexFDO

Authors Eliza O'reilly, Venkat Chandrasekaran

Publication date 7 June 2023

Published in SIAM Journal on Optimization (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2110.14779

zbMATH Keywords

approximation of convex bodiesconvex regressionsemidefinite programmingsupport function estimation

Mathematics Subject Classification ID

Nonparametric regression and quantile regression (62G08) Asymptotic properties of nonparametric inference (62G20) Semidefinite programming (90C22) Minimax problems in mathematical programming (90C47)

Abstract: Convex regression is the problem of fitting a convex function to a data set consisting of input-output pairs. We present a new approach to this problem called spectrahedral regression, in which we fit a spectrahedral function to the data, i.e. a function that is the maximum eigenvalue of an affine matrix expression of the input. This method represents a significant generalization of polyhedral (also called max-affine) regression, in which a polyhedral function (a maximum of a fixed number of affine functions) is fit to the data. We prove bounds on how well spectrahedral functions can approximate arbitrary convex functions via statistical risk analysis. We also analyze an alternating minimization algorithm for the non-convex optimization problem of fitting the best spectrahedral function to a given data set. We show that this algorithm converges geometrically with high probability to a small ball around the optimal parameter given a good initialization. Finally, we demonstrate the utility of our approach with experiments on synthetic data sets as well as real data arising in applications such as economics and engineering design.

Recommendations

• Consistency of multidimensional convex regression
• A dual active set algorithm for optimal sparse convex regression
• A Computational Framework for Multivariate Convex Regression and Its Variants
• Multivariate convex regression with adaptive partitioning
• Convex piecewise-linear fitting

Cites work

• A useful variant of the Davis-Kahan theorem for statisticians
• An introduction to matrix concentration inequalities
• Approximation of convex sets by polytopes
• Composite difference-MAX programs for modern statistical estimation problems
• Convex piecewise-linear fitting
• Data fitting with geometric-programming-compatible softmax functions
• Fitting tractable convex sets to support function evaluations
• Hanson-Wright inequality and sub-Gaussian concentration
• High-dimensional probability. An introduction with applications in data science
• Lifting for simplicity: concise descriptions of convex sets
• Max-Affine Regression: Parameter Estimation for Gaussian Designs
• Metric entropy of some classes of sets with differentiable boundaries
• Multivariate convex regression with adaptive partitioning
• Nonparametric least squares estimation of a multivariate convex regression function
• Optimal rates of convergence for convex set estimation from support functions
• Probability in Banach spaces. Isoperimetry and processes
• Small ball probabilities for linear images of high-dimensional distributions
• Small ball probability estimates for log-concave measures
• Uniform Hanson-Wright type concentration inequalities for unbounded entries via the entropy method
• Weak convergence and empirical processes. With applications to statistics

Cited in

(3)

• Max-affine regression via first-order methods
• Learning elastoplasticity with implicit layers
• Spectral methods for polynomial optimization

This page was built for publication: Spectrahedral Regression

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6155881)