A Hybrid Causal Search Algorithm for Latent Variable Models

Juan Miguel Ogarrio, Peter Spirtes, Joe Ramsey
Proceedings of the Eighth International Conference on Probabilistic Graphical Models, PMLR 52:368-379, 2016.

Abstract

Existing score-based causal model search algorithms such as \textitGES (and a speeded up version, \textitFGS) are asymptotically correct, fast, and reliable, but make the unrealistic assumption that the true causal graph does not contain any unmeasured confounders. There are several constraint-based causal search algorithms (e.g \textitRFCI, \emphFCI, or \emphFCI+) that are asymptotically correct without assuming that there are no unmeasured confounders, but often perform poorly on small samples. We describe a combined score and constraint-based algorithm, \emphGFCI, that we prove is asymptotically correct. On synthetic data, \textitGFCI is only slightly slower than \emphRFCI but more accurate than \textitFCI, \textitRFCI and \textitFCI+.

Cite this Paper

BibTeX
@InProceedings{pmlr-v52-ogarrio16, title = {A Hybrid Causal Search Algorithm for Latent Variable Models}, author = {Ogarrio, Juan Miguel and Spirtes, Peter and Ramsey, Joe}, booktitle = {Proceedings of the Eighth International Conference on Probabilistic Graphical Models}, pages = {368--379}, year = {2016}, editor = {Antonucci, Alessandro and Corani, Giorgio and Campos}, Cassio Polpo}, volume = {52}, series = {Proceedings of Machine Learning Research}, address = {Lugano, Switzerland}, month = {06--09 Sep}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v52/ogarrio16.pdf}, url = {https://proceedings.mlr.press/v52/ogarrio16.html}, abstract = {Existing score-based causal model search algorithms such as \textitGES (and a speeded up version, \textitFGS) are asymptotically correct, fast, and reliable, but make the unrealistic assumption that the true causal graph does not contain any unmeasured confounders. There are several constraint-based causal search algorithms (e.g \textitRFCI, \emphFCI, or \emphFCI+) that are asymptotically correct without assuming that there are no unmeasured confounders, but often perform poorly on small samples. We describe a combined score and constraint-based algorithm, \emphGFCI, that we prove is asymptotically correct. On synthetic data, \textitGFCI is only slightly slower than \emphRFCI but more accurate than \textitFCI, \textitRFCI and \textitFCI+.} }
Endnote
%0 Conference Paper %T A Hybrid Causal Search Algorithm for Latent Variable Models %A Juan Miguel Ogarrio %A Peter Spirtes %A Joe Ramsey %B Proceedings of the Eighth International Conference on Probabilistic Graphical Models %C Proceedings of Machine Learning Research %D 2016 %E Alessandro Antonucci %E Giorgio Corani %E Cassio Polpo Campos} %F pmlr-v52-ogarrio16 %I PMLR %P 368--379 %U https://proceedings.mlr.press/v52/ogarrio16.html %V 52 %X Existing score-based causal model search algorithms such as \textitGES (and a speeded up version, \textitFGS) are asymptotically correct, fast, and reliable, but make the unrealistic assumption that the true causal graph does not contain any unmeasured confounders. There are several constraint-based causal search algorithms (e.g \textitRFCI, \emphFCI, or \emphFCI+) that are asymptotically correct without assuming that there are no unmeasured confounders, but often perform poorly on small samples. We describe a combined score and constraint-based algorithm, \emphGFCI, that we prove is asymptotically correct. On synthetic data, \textitGFCI is only slightly slower than \emphRFCI but more accurate than \textitFCI, \textitRFCI and \textitFCI+.
RIS
TY - CPAPER TI - A Hybrid Causal Search Algorithm for Latent Variable Models AU - Juan Miguel Ogarrio AU - Peter Spirtes AU - Joe Ramsey BT - Proceedings of the Eighth International Conference on Probabilistic Graphical Models DA - 2016/08/15 ED - Alessandro Antonucci ED - Giorgio Corani ED - Cassio Polpo Campos} ID - pmlr-v52-ogarrio16 PB - PMLR DP - Proceedings of Machine Learning Research VL - 52 SP - 368 EP - 379 L1 - http://proceedings.mlr.press/v52/ogarrio16.pdf UR - https://proceedings.mlr.press/v52/ogarrio16.html AB - Existing score-based causal model search algorithms such as \textitGES (and a speeded up version, \textitFGS) are asymptotically correct, fast, and reliable, but make the unrealistic assumption that the true causal graph does not contain any unmeasured confounders. There are several constraint-based causal search algorithms (e.g \textitRFCI, \emphFCI, or \emphFCI+) that are asymptotically correct without assuming that there are no unmeasured confounders, but often perform poorly on small samples. We describe a combined score and constraint-based algorithm, \emphGFCI, that we prove is asymptotically correct. On synthetic data, \textitGFCI is only slightly slower than \emphRFCI but more accurate than \textitFCI, \textitRFCI and \textitFCI+. ER -
APA
Ogarrio, J.M., Spirtes, P. & Ramsey, J.. (2016). A Hybrid Causal Search Algorithm for Latent Variable Models. Proceedings of the Eighth International Conference on Probabilistic Graphical Models, in Proceedings of Machine Learning Research 52:368-379 Available from https://proceedings.mlr.press/v52/ogarrio16.html.

Related Material