Confounder detection in high-dimensional linear models using first moments of spectral measures

DOI10.1162/NECO_A_01099MaRDI QIDQ5157228zbMATH OpenOpenAlexWikidataFDO

Authors Furui Liu, Lai Wan Chan

Publication date 12 October 2021

Published in Neural Computation (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1803.06852

zbMATH Keywords

spectral measure

Mathematics Subject Classification ID

Linear regression; mixed models (62J05) Inference from stochastic processes and spectral analysis (62M15)

Abstract: In this paper, we study the confounder detection problem in the linear model, where the target variable $Y$ is predicted using its $n$ potential causes $X_n=(x_1,\mathrm{...},x_n{)}^T$. Based on an assumption of rotation invariant generating process of the model, recent study shows that the spectral measure induced by the regression coefficient vector with respect to the covariance matrix of $X_n$ is close to a uniform measure in purely causal cases, but it differs from a uniform measure characteristically in the presence of a scalar confounder. Then, analyzing spectral measure pattern could help to detect confounding. In this paper, we propose to use the first moment of the spectral measure for confounder detection. We calculate the first moment of the regression vector induced spectral measure, and compare it with the first moment of a uniform spectral measure, both defined with respect to the covariance matrix of $X_n$. The two moments coincide in non-confounding cases, and differ from each other in the presence of confounding. This statistical causal-confounding asymmetry can be used for confounder detection. Without the need of analyzing the spectral measure pattern, our method does avoid the difficulty of metric choice and multiple parameter optimization. Experiments on synthetic and real data show the performance of this method.

Recommendations

• Detecting multiple confounders
• On the definition of a confounder
• Identifying causal effects with proxy variables of an unmeasured confounder
• A new criterion for confounder selection
• Nonlinear confounding in high-dimensional regression

Cites work

• A linear non-Gaussian acyclic model for causal discovery
• Bounding \(\bar d\)-distance by informational divergence: A method to prove measure concentration
• Causal Inference Using the Algorithmic Markov Condition
• Causal Inference on Discrete Data via Estimating Distance Correlations
• Concentration of measure and isoperimetric inequalities in product spaces
• DirectLiNGAM: a direct method for learning a linear non-Gaussian structural equation model
• Estimation of a structural vector autoregression model using non-Gaussianity
• Estimation of causal effects using linear non-Gaussian causal models with hidden variables
• Pairwise likelihood ratios for estimation of non-Gaussian structural equation models
• Random polynomials of high degree and Lévy concentration of measure
• Rank-one modification of the symmetric eigenproblem

Cited in

(2)

• The blessings of multiple causes
• A consistent estimator for confounding strength

Describes a project that uses

Uses Software

• DirectLiNGAM

This page was built for publication: Confounder detection in high-dimensional linear models using first moments of spectral measures

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5157228)