Data-driven covariate selection for nonparametric estimation of causal effects

Doris Entner, Patrik Hoyer, Peter Spirtes
Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics, PMLR 31:256-264, 2013.

Abstract

The estimation of causal effects from non-experimental data is a fundamental problem in many fields of science. One of the main obstacles concerns confounding by observed or latent covariates, an issue which is typically tackled by adjusting for some set of observed covariates. In this contribution, we analyze the problem of inferring whether a given variable has a causal effect on another and, if it does, inferring an adjustment set of covariates that yields a consistent and unbiased estimator of this effect, based on the (conditional) independence and dependence relationships among the observed variables. We provide two elementary rules that we show to be both sound and complete for this task, and compare the performance of a straightforward application of these rules with standard alternative procedures for selecting adjustment sets.

Cite this Paper

BibTeX
@InProceedings{pmlr-v31-entner13a, title = {Data-driven covariate selection for nonparametric estimation of causal effects}, author = {Entner, Doris and Hoyer, Patrik and Spirtes, Peter}, booktitle = {Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics}, pages = {256--264}, year = {2013}, editor = {Carvalho, Carlos M. and Ravikumar, Pradeep}, volume = {31}, series = {Proceedings of Machine Learning Research}, address = {Scottsdale, Arizona, USA}, month = {29 Apr--01 May}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v31/entner13a.pdf}, url = {https://proceedings.mlr.press/v31/entner13a.html}, abstract = {The estimation of causal effects from non-experimental data is a fundamental problem in many fields of science. One of the main obstacles concerns confounding by observed or latent covariates, an issue which is typically tackled by adjusting for some set of observed covariates. In this contribution, we analyze the problem of inferring whether a given variable has a causal effect on another and, if it does, inferring an adjustment set of covariates that yields a consistent and unbiased estimator of this effect, based on the (conditional) independence and dependence relationships among the observed variables. We provide two elementary rules that we show to be both sound and complete for this task, and compare the performance of a straightforward application of these rules with standard alternative procedures for selecting adjustment sets.} }
Endnote
%0 Conference Paper %T Data-driven covariate selection for nonparametric estimation of causal effects %A Doris Entner %A Patrik Hoyer %A Peter Spirtes %B Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2013 %E Carlos M. Carvalho %E Pradeep Ravikumar %F pmlr-v31-entner13a %I PMLR %P 256--264 %U https://proceedings.mlr.press/v31/entner13a.html %V 31 %X The estimation of causal effects from non-experimental data is a fundamental problem in many fields of science. One of the main obstacles concerns confounding by observed or latent covariates, an issue which is typically tackled by adjusting for some set of observed covariates. In this contribution, we analyze the problem of inferring whether a given variable has a causal effect on another and, if it does, inferring an adjustment set of covariates that yields a consistent and unbiased estimator of this effect, based on the (conditional) independence and dependence relationships among the observed variables. We provide two elementary rules that we show to be both sound and complete for this task, and compare the performance of a straightforward application of these rules with standard alternative procedures for selecting adjustment sets.
RIS
TY - CPAPER TI - Data-driven covariate selection for nonparametric estimation of causal effects AU - Doris Entner AU - Patrik Hoyer AU - Peter Spirtes BT - Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics DA - 2013/04/29 ED - Carlos M. Carvalho ED - Pradeep Ravikumar ID - pmlr-v31-entner13a PB - PMLR DP - Proceedings of Machine Learning Research VL - 31 SP - 256 EP - 264 L1 - http://proceedings.mlr.press/v31/entner13a.pdf UR - https://proceedings.mlr.press/v31/entner13a.html AB - The estimation of causal effects from non-experimental data is a fundamental problem in many fields of science. One of the main obstacles concerns confounding by observed or latent covariates, an issue which is typically tackled by adjusting for some set of observed covariates. In this contribution, we analyze the problem of inferring whether a given variable has a causal effect on another and, if it does, inferring an adjustment set of covariates that yields a consistent and unbiased estimator of this effect, based on the (conditional) independence and dependence relationships among the observed variables. We provide two elementary rules that we show to be both sound and complete for this task, and compare the performance of a straightforward application of these rules with standard alternative procedures for selecting adjustment sets. ER -
APA
Entner, D., Hoyer, P. & Spirtes, P.. (2013). Data-driven covariate selection for nonparametric estimation of causal effects. Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 31:256-264 Available from https://proceedings.mlr.press/v31/entner13a.html.

Related Material