Structure Learning for Extremal Tree Models
Authors:
- 1 November 2022
- journal article
- research article
- Published by Oxford University Press (OUP) in Journal of the Royal Statistical Society Series B: Statistical Methodology
- Vol. 84 (5) , 2055-2087
Abstract
Extremal graphical models are sparse statistical models for multivariate extreme events. The underlying graph encodes conditional independencies and enables a visual interpretation of the complex extremal dependence structure. For the important case of tree models, we develop a data-driven methodology for learning the graphical structure. We show that sample versions of the extremal correlation and a new summary statistic, which we call the extremal variogram, can be used as weights for a minimum spanning tree to consistently recover the true underlying tree. Remarkably, this implies that extremal tree models can be learned in a completely non-parametric fashion by using simple summary statistics and without the need to assume discrete distributions, existence of densities or parametric models for bivariate distributions.Keywords
All Related Versions
Funding Information
- Natural Sciences and Engineering Research Council of Canada
- Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung (161297)
This publication has 34 references indexed in Scilit:
- Conditional simulation of max-stable processesBiometrika, 2012
- An M-estimator for tail dependence in arbitrary dimensionsThe Annals of Statistics, 2012
- Extremal dependence measure and extremogram: the regularly varying caseExtremes, 2011
- Multivariate generalized Pareto distributionsBernoulli, 2006
- A dependence measure for multivariate and spatial extreme values: Properties and inferenceBiometrika, 2003
- Statistics of extremes in hydrologyAdvances in Water Resources, 2002
- ARCH modeling in financeJournal of Econometrics, 1992
- Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of United Kingdom InflationEconometrica, 1982
- Approximating discrete probability distributions with dependence treesIEEE Transactions on Information Theory, 1968
- On the shortest spanning subtree of a graph and the traveling salesman problemProceedings of the American Mathematical Society, 1956