Abstract
Learning causal models hidden in the background of observational data has been a difficult issue. Dealing with latent common causes and selection bias for constructing causal models in real data is often necessary because observing all relevant variables is difficult. Ancestral graph models are effective and useful for representing causal models with some information of such latent variables. The causal faithfulness condition, which is usually assumed for determining the models, is known to often be weakly violated in statistical view points for finite data. One of the authors developed a constraint-based causal learning algorithm that is robust against the weak violations while assuming no latent variables. In this study, we applied and extended the thoughts of the algorithm to the inference of ancestral graph models. The practical validity and effectiveness of the algorithm are also confirmed by using some standard datasets in comparison with FCI and RFCI algorithms.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.Notes
In this paper, we assume that a v-structure is interpreted as a causal v-structure; thus we omit the term causal in the structure.
We use the term weak in this paper because we cannot recover causal graphical models from observational data if the CFC strongly violates.
About Possible-D-Sep, refer to Spirtes et al. [19].
The datasets are available from http://www.cs.huji.ac.il/site/labs/compbio/Repository/.
References
Abramson, B., Brown, J., Winkler, R.L.: Hailfinder: a Bayesian system for forecasting severe weather. Int. J. Forecast. 12, 57–71 (1996)
Beinlich, I., Suermondt, H., Chavez, R., Cooper, G.: The ALARM monitoring system: A case study with two probabilistic inference techniques for belief networks. In: Proc. of European Conference on Artificial Intelligence in Medicine (AIME-89), pp. 247–256 (1989)
Binder, J., Koller, D., Russell, S., Kanazawa, K.: Adaptive probabilistic networks with hidden variables. Mach. Learn. 29, 213–244 (1997)
Claassen, T., Heskes, T.: A Bayesian approach to constraint based causal inference. In: Proc. of Conference on Uncertainty in Artificial Intelligence (UAI-12), pp. 207–216 (2012)
Colombo, D., Maathuis, M.H., Kalisch, M., Richardson, T.S.: Learning high-dimensional directed acyclic graphs with latent and selection variables. Ann. Stat. 40(1), 294–321 (2012)
Dawid, A.P.: Conditional independence for statistical operations. Ann. Stat. 8, 598–617 (1980)
Geiger, D., Verma, T., Pearl, J.: d-separation: from theorems to algorithms. In: Proc. of Conference on Uncertainty in Artificial Intelligence (UAI-89), pp. 139–148 (1989)
Isozaki, T.: A robust causal discovery algorithm against faithfulness violation. Trans. Jpn. Soc. Artif. Intell. 29(1), 137–147 (2014)
Isozaki, T., Kuroki, M.: Learning maximal ancestral graphs with robustness for faithfulness violations. In: Proc. of Second International Workshop on Advanced Methodologies for Bayesian Networks (AMBN-15), pp. 196–208 (2016)
Jensen, F.V., Kjærulff, U., Olesen, G., Pedersen, J.: An expert system for control of waste water treatment. Technical report, Judex Datasystemer A/S, Aalborg, Denmark (1989, in Danish)
Kalisch, M., Mächler, M., Colombo, D., Maathuis, M.H., Bühlmann, P.: Causal inference using graphical models with the R package pcalg. J. Stat. Softw. 47(11), 1–26 (2012)
Meek, C.: Strong completeness and faithfulness in Bayesian networks. In: Proc. of Conference on Uncertainty in Artificial Intelligence (UAI-95), pp. 411–418 (1995)
Pearl, J.: Probabilistic Reasoning in Intelligent Systems. Morgan Kaufmann, San Mateo (1988)
Pearl, J., Verma, T.: A theory of inferred causation. In: Proc. of Conference on Principles of Knowledge Representation and Reasoning, pp. 441–452 (1991)
Pellet, J., Elisseeff, A.: Finding latent causes in causal networks: an efficient approach based on markov blankets. In: Proc. of Advances in Neural Information Processing Systems 21 (NIPS 08), pp. 1249–1256 (2008)
Ramsey, J., Spirtes, P., Zhang, J.: Adjacency-faithfulness and conservative causal inference. In: Proc. of Conference on Uncertainty in Artificial Intelligence (UAI-06), pp. 401–408 (2006)
Reichenbach, H.: The Direction of Time. Dover Publications, Mineola (1956, Republication of the work published by University of California Press, Berkely)
Richardson, T., Spirtes, P.: Ancestral graph markov models. Ann. Stat. 30(4), 962–1030 (2002)
Spirtes, P., Glymour, C., Scheines, R.: Causation, Prediction and Search, 2nd edn. MIT Press, Cambridge (2000)
Spirtes, P., Meek, C., Richardson, T.: Causal inference in the presence of latent variables and selection bias. In: Proc. of Conference on Uncertainty in Artificial Intelligence (UAI-95), pp. 499–506 (1995)
Williamson, J.: Bayesian Nets and Causality: Philosophical and Computational Foundations. Oxford University Press, Oxford (2005)
Xie, X., Geng, Z.: A recursive method for structural learning of directed acyclic graphs. J. Mach. Learn. Res. 9, 459–483 (2008)
Yehezkel, R., Lerner, B.: Bayesian network structure learning by recursive autonomy identification. J. Mach. Learn. Res. 10, 1527–1570 (2009)
Zhang, J.: On the completeness of orientation rules for causal discovery in the presence of latent confounders and selection bias. Artif. Intell. 172, 1873–1896 (2008)
Zhang, J., Spirtes, P.: Detection of unfaithfulness and robust causal inference. Minds Mach. 18(2), 239–271 (2008)
Acknowledgements
The authors thank Tomohide Haruguchi for assisting with the experiments. Isozaki thanks Hiroaki Kitano of Sony Computer Science Laboratories, Inc. for his support. The authors would like to thank the anonymous reviewers for their helpful and constructive comments, which contributed to improving the paper.
Author information
Authors and Affiliations
Corresponding author
About this article
Cite this article
Isozaki, T., Kuroki, M. Learning Causal Graphs with Latent Confounders in Weak Faithfulness Violations. New Gener. Comput. 35, 29–45 (2017). https://doi.org/10.1007/s00354-016-0003-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00354-016-0003-x