An efficient causal discovery algorithm for linear models

Bayesian network learning algorithms have been widely used for causal discovery since the pioneer work [13,18]. Among all existing algorithms, three-phase dependency analysis algorithm (TPDA) [5] is the most efficient one in the sense that it has polynomial-time complexity. However, there are still some limitations to be improved. First, TPDA depends on mutual information-based conditional independence (CI) tests, and so is not easy to be applied to continuous data. In addition, TPDA uses two phases to get approximate skeletons of Bayesian networks, which is not efficient in practice. In this paper, we propose a two-phase algorithm with partial correlation-based CI tests: the first phase of the algorithm constructs a Markov random field from data, which provides a close approximation to the structure of the true Bayesian network; at the second phase, the algorithm removes redundant edges according to CI tests to get the true Bayesian network. We show that two-phase algorithm with partial correlation-based CI tests can deal with continuous data following arbitrary distributions rather than only Gaussian distribution.

[1]  David A. Bell,et al.  Learning Bayesian networks from data: An information-theory based approach , 2002, Artif. Intell..

[2]  André Elisseeff,et al.  A Partial Correlation-Based Algorithm for Causal Structure Discovery with Continuous Variables , 2007, IDA.

[3]  Korbinian Strimmer,et al.  From correlation to causation networks: a simple approximate learning algorithm and its application to high-dimensional plant gene expression data , 2007, BMC Systems Biology.

[4]  P. Spirtes,et al.  An Algorithm for Fast Recovery of Sparse Causal Graphs , 1991 .

[5]  Tom Burr,et al.  Causation, Prediction, and Search , 2003, Technometrics.

[6]  Stuart J. Russell,et al.  Adaptive Probabilistic Networks with Hidden Variables , 1997, Machine Learning.

[7]  Aapo Hyvärinen,et al.  Causal discovery of linear acyclic models with arbitrary distributions , 2008, UAI.

[8]  Judea Pearl,et al.  A Theory of Inferred Causation , 1991, KR.

[9]  Aapo Hyvärinen,et al.  Causal modelling combining instantaneous and lagged effects: an identifiable model based on non-Gaussianity , 2008, ICML '08.

[10]  Franz von Kutschera,et al.  Causation , 1993, J. Philos. Log..

[11]  J. Pearl.  Causality: Models, Reasoning and Inference , 2000 .

[12]  Kristian Kristensen,et al.  The use of a Bayesian network in the design of a decision support system for growing malting barley without use of pesticides , 2002 .

[13]  Lai-Wan Chan,et al.  A Heuristic Partial-Correlation-Based Algorithm for Causal Relationship Discovery on Continuous Data , 2009, IDEAL.

[14]  R. Shibata,et al.  PARTIAL CORRELATION AND CONDITIONAL CORRELATION AS MEASURES OF CONDITIONAL INDEPENDENCE , 2004 .

[15]  J. Laurie Snell,et al.  Markov Random Fields and Their Applications , 1980 .

[16]  Aapo Hyvärinen,et al.  A Linear Non-Gaussian Acyclic Model for Causal Discovery , 2006, J. Mach. Learn. Res..

[17]  P. Spirtes,et al.  From probability to causality , 1991 .

[18]  Allan Leck Jensen,et al.  MIDAS: An Influence Diagram for Management of Mildew in Winter Wheat , 1996, UAI.

[19]  S Andreassen,et al.  Evaluation of the diagnostic performance of the expert EMG assistant MUNIN. , 1996, Electroencephalography and clinical neurophysiology.

[20]  Gregory F. Cooper,et al.  The ALARM Monitoring System: A Case Study with two Probabilistic Inference Techniques for Belief Networks , 1989, AIME.