Detecting confounding in multivariate linear models via spectral analysis

04/05/2017
by   Dominik Janzing, et al.
0

We study a model where one target variable Y is correlated with a vector X:=(X_1,...,X_d) of predictor variables being potential causes of Y. We describe a method that infers to what extent the statistical dependences between X and Y are due to the influence of X on Y and to what extent due to a hidden common cause (confounder) of X and Y. The method relies on concentration of measure results for large dimensions d and an independence assumption stating that, in the absence of confounding, the vector of regression coefficients describing the influence of each X on Y typically has `generic orientation' relative to the eigenspaces of the covariance matrix of X. For the special case of a scalar confounder we show that confounding typically spoils this generic orientation in a characteristic way that can be used to quantitatively estimate the amount of confounding.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
03/02/2018

Detecting non-causal artifacts in multivariate linear regression models

We consider linear models where d potential causes X_1,...,X_d are corre...
research
03/19/2018

Confounder Detection in High Dimensional Linear Models using First Moments of Spectral Measures

In this paper, we study the confounder detection problem in the linear m...
research
11/03/2022

A Consistent Estimator for Confounding Strength

Regression on observational data can fail to capture a causal relationsh...
research
04/08/2020

Doubly Debiased Lasso: High-Dimensional Inference under Hidden Confounding and Measurement Errors

Inferring causal relationships or related associations from observationa...
research
11/13/2018

Spectral Deconfounding and Perturbed Sparse Linear Models

Standard high-dimensional regression methods assume that the underlying ...
research
08/11/2019

A new Granger causality measure for eliminating the confounding influence of latent common inputs

In this paper, we propose a new Granger causality measure which is robus...
research
02/04/2022

Correcting Confounding via Random Selection of Background Variables

We propose a method to distinguish causal influence from hidden confound...

Please sign up or login with your details

Forgot password? Click here to reset