Elliptical graphical modelling
Preprint
Abstract
We propose elliptical graphical models based on conditional uncorrelatedness as a general- ization of Gaussian graphical models by letting the population distribution be elliptical instead of normal, allowing the fitting of data with arbitrarily heavy tails. We study the class of propor- tionally affine equivariant scatter estimators and show how they can be used to perform elliptical graphical modelling, leading to a new class of partial correlation estimators and analogues of the classical deviance test. General expressions for the asymptotic variance of partial correla- tion estimators, unconstrained and under decomposable models, are given, and the asymptotic chi square approximation of the pseudo-deviance test statistic is proved. The feasibility of our approach is demonstrated by a simulation study, using, among others, Tyler's scatter estimator, which is distribution-free within the elliptical model. Our approach provides a robustification of Gaussian graphical modelling. The latter is likelihood-based and known to be very sensitive to model misspecification and outlying observations.Keywords
All Related Versions
- Version 1, 2015-06-13, ArXiv
- Published version: Biometrika, 98 (4), 935.
This publication has 16 references indexed in Scilit:
- Robust concentration graph model selectionComputational Statistics & Data Analysis, 2010
- A note on multivariate location and scatter statistics for sparse data setsStatistics & Probability Letters, 2010
- A canonical definition of shapeStatistics & Probability Letters, 2008
- Sparse inverse covariance estimation with the graphical lassoBiostatistics, 2007
- Robust Gaussian graphical modelingJournal of Multivariate Analysis, 2006
- PARTIAL CORRELATION AND CONDITIONAL CORRELATION AS MEASURES OF CONDITIONAL INDEPENDENCEAustralian & New Zealand Journal of Statistics, 2004
- The affine equivariant sign covariance matrix: asymptotic behavior and efficienciesJournal of Multivariate Analysis, 2003
- A practical affine equivariant multivariate medianBiometrika, 2002
- Influence Function and Efficiency of the Minimum Covariance Determinant Scatter Matrix EstimatorJournal of Multivariate Analysis, 1999
- On Tyler's M-Functional of Scatter in High DimensionAnnals of the Institute of Statistical Mathematics, 1998