A Linear Non-Gaussian Acyclic Model for Causal Discovery

In recent years, several methods have been proposed for the discovery of causal structure from non-experimental data. Such methods make various assumptions on the data generating process to facilitate its identification from purely observational data. Continuing this line of research, we show how to discover the complete causal structure of continuous-valued data, under the assumptions that (a) the data generating process is linear, (b) there are no unobserved confounders, and (c) disturbance variables have non-Gaussian distributions of non-zero variances. The solution relies on the use of the statistical method known as independent component analysis, and does not require any pre-specified time-ordering of the variables. We provide a complete Matlab package for performing this LiNGAM analysis (short for Linear Non-Gaussian Acyclic Model), and demonstrate the effectiveness of the method using artificially generated data and real-world data.

[1]  David Heckerman,et al.  Learning Gaussian Networks , 1994, UAI.

[2]  A. Tamhane,et al.  Multiple Comparison Procedures , 1989 .

[3]  Y. Hochberg.  A sharper Bonferroni procedure for multiple tests of significance , 1988 .

[4]  Aapo Hyvärinen,et al.  Finding a causal ordering via independent component analysis , 2006, Comput. Stat. Data Anal..

[5]  M. Kenward,et al.  An Introduction to the Bootstrap , 2007 .

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

[7]  Motoaki Kawanabe,et al.  Estimating Functions for Blind Separation When Sources Have Variance Dependencies , 2004, J. Mach. Learn. Res..

[8]  A. Hyvärinen,et al.  One-unit contrast functions for independent component analysis: a statistical analysis , 1997 .

[9]  Pierre Comon.  Independent component analysis - a new concept? signal processing , 1994 .

[10]  Shohei Shimizu,et al.  Use of non-normality in structural equation modeling: Application to direction of causation , 2008 .

[11]  Ke-Hai Yuan,et al.  Mean and Covariance Structure Analysis: Theoretical and Practical Improvements , 1997 .

[12]  Richard A. Davis,et al.  Time Series: Theory and Methods , 2013 .

[13]  Patrik O. Hoyer,et al.  Estimation of linear, non-gaussian causal models in the presence of confounding latent variables , 2006, Probabilistic Graphical Models.

[14]  Aapo Hyvärinen,et al.  Discovery of Non-gaussian Linear Causal Models using ICA , 2005, UAI.

[15]  J. Cardoso,et al.  Blind beamforming for non-gaussian signals , 1993 .

[16]  P. Young,et al.  Time series analysis, forecasting and control , 1972, IEEE Transactions on Automatic Control.

[17]  Y. Benjamini,et al.  Controlling the false discovery rate: a practical and powerful approach to multiple testing , 1995 .

[18]  A. Adelstein.  National statistics. , 1975, Postgraduate medical journal.

[19]  Jean-François Cardoso,et al.  Equivariant adaptive source separation , 1996, IEEE Trans. Signal Process..

[20]  Y Kano,et al.  Can test statistics in covariance structure analysis be trusted? , 1992, Psychological bulletin.

[21]  George E. P. Box,et al.  Time Series Analysis: Forecasting and Control , 1977 .

[22]  R. Simes,et al.  An improved Bonferroni procedure for multiple tests of significance , 1986 .

[23]  Pierre Comon,et al.  Independent component analysis, A new concept? , 1994, Signal Process..

[24]  Erkki Oja,et al.  Performance analysis of the FastICA algorithm and Crame/spl acute/r-rao bounds for linear independent component analysis , 2006, IEEE Transactions on Signal Processing.

[25]  M. Browne.  Asymptotically distribution-free methods for the analysis of covariance structures. , 1984, The British journal of mathematical and statistical psychology.

[26]  Aapo Hyvärinen,et al.  Validating the independent components of neuroimaging time series via clustering and visualization , 2004, NeuroImage.

[27]  P. Spirtes,et al.  Causation, Prediction, and Search, 2nd Edition , 2001 .

[28]  H. Keselman,et al.  Multiple Comparison Procedures , 2005 .

[29]  E. Oja,et al.  Corrections to "Performance Analysis of the FastICA Algorithm and Cram´ er-Rao Bounds for Linear Independent Component Analysis", TSP 04/06 , 2007 .

[30]  Y. Dodge,et al.  On Asymmetric Properties of the Correlation Coeffcient in the Regression Setting , 2001 .

[31]  Rainer E. Burkard,et al.  Linear Assignment Problems and Extensions , 1999, Handbook of Combinatorial Optimization.

[32]  S. Holm.  A Simple Sequentially Rejective Multiple Test Procedure , 1979 .

[33]  Aapo Hyvärinen,et al.  Testing Significance of Mixing and Demixing Coefficients in ICA , 2006, ICA.

[34]  Philippe Garat,et al.  Blind separation of mixture of independent sources through a quasi-maximum likelihood approach , 1997, IEEE Trans. Signal Process..

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

[36]  Aapo Hyvärinen,et al.  Fast and robust fixed-point algorithms for independent component analysis , 1999, IEEE Trans. Neural Networks.

[37]  Aapo Hyvärinen,et al.  New Permutation Algorithms for Causal Discovery Using ICA , 2006, ICA.