On the estimation of nested Archimedean copulas: a theoretical and an experimental comparison

A lot of progress regarding estimation of nested Archimedean copulas has been booked since their introduction by Joe (Multivariate models and dependence concepts. Chapman and Hall, London, 1997). The currently published procedures can be seen as particular cases of two different, more general, approaches. In the first approach, the tree structure of the target nested Archimedean copulas is estimated using hierarchical clustering to get a binary tree, and then parts of this binary tree are collapsed according to some strategy. This two-step estimation of the tree structure paves the way for estimation of the generators according to the sufficient nesting condition afterwards, this sufficient nesting condition on the generators ensuring the resulting estimated nested Archimedean copula is a proper copula. In contrast to the first approach, the second approach estimates the tree structure free of any concern for the generators. While this is the main strength of this second approach, it is also its main weakness: estimation of the generators afterwards so that the resulting nested Archimedean copula is a proper copula still lacks a solution. In this paper, both approaches are formally explored, detailed explanations and examples are given, as well as results from a performance study where a new way of comparing tree structure estimators is offered. A nested Archimedean copula is also estimated based on exams results from 482 students, and a naive attempt to check the fit is made using principal component analysis.

[1]  Nicholas C. Wormald,et al.  Reconstruction of Rooted Trees From Subtrees , 1996, Discret. Appl. Math..

[2]  K. Nixon The Parsimony Ratchet, a New Method for Rapid Parsimony Analysis , 1999 .

[3]  M. Rezapour On the construction of nested Archimedean copulas for d-monotone generators , 2015 .

[4]  Jeffrey Sheen,et al.  Conditional Systemic Risk with Penalized Copula , 2015 .

[5]  Johan Segers,et al.  Nonparametric estimation of the tree structure of a nested Archimedean copula , 2013, Comput. Stat. Data Anal..

[6]  O. Bininda-Emonds,et al.  The evolution of supertrees. , 2004, Trends in ecology & evolution.

[7]  Martin Holena,et al.  Using Copulas in Data Mining Based on the Observational Calculus , 2015, IEEE Transactions on Knowledge and Data Engineering.

[8]  Oliver Eulenstein,et al.  The shape of supertrees to come: tree shape related properties of fourteen supertree methods. , 2005, Systematic biology.

[9]  Pablo N. Hess,et al.  An empirical test of the midpoint rooting method , 2007, Biological journal of the Linnean Society. Linnean Society of London.

[10]  Ostap Okhrin,et al.  Hierarchical Archimedean Copulae: The HAC Package , 2012 .

[11]  Yarema Okhrin,et al.  On the structure and estimation of hierarchical Archimedean copulas , 2013 .

[12]  N. Saitou,et al.  The neighbor-joining method: a new method for reconstructing phylogenetic trees. , 1987, Molecular biology and evolution.

[13]  Stefano Ferilli A sentence structure-based approach to unsupervised author identification , 2014, Journal of Intelligent Information Systems.

[14]  F J Ayala,et al.  Tree rooting with outgroups when they differ in their nucleotide composition from the ingroup: the Drosophila saltans and willistoni groups, a case study. , 2000, Molecular phylogenetics and evolution.

[15]  W. Hoeffding A Non-Parametric Test of Independence , 1948 .

[16]  Yarema Okhrin,et al.  Properties of hierarchical Archimedean copulas , 2013 .

[17]  A. McNeil Sampling nested Archimedean copulas , 2008 .

[18]  R. Tibshirani,et al.  Regularisation Paths for Conditional Logistic Regression: the clogitL1 package , 2014, 1405.3344.

[19]  P. Filzmoser,et al.  Robust Maximum Association Estimators , 2017 .

[20]  Alexander J. McNeil,et al.  Multivariate Archimedean copulas, $d$-monotone functions and $\ell_1$-norm symmetric distributions , 2009, 0908.3750.

[21]  Marius Hofert,et al.  Densities of nested Archimedean copulas , 2012, J. Multivar. Anal..

[22]  Marius Hofert,et al.  Nested Archimedean Copulas Meet R: The nacopula Package , 2011 .

[23]  Liam J. Revell,et al.  phytools: an R package for phylogenetic comparative biology (and other things) , 2012 .

[24]  Martin Holena,et al.  An approach to structure determination and estimation of hierarchical Archimedean Copulas and its application to Bayesian classification , 2016, Journal of Intelligent Information Systems.

[25]  H. Joe Multivariate models and dependence concepts , 1998 .

[26]  Ward C. Wheeler,et al.  NUCLEIC ACID SEQUENCE PHYLOGENY AND RANDOM OUTGROUPS , 1990, Cladistics : the international journal of the Willi Hennig Society.

[27]  Marius Hofert,et al.  Efficiently sampling nested Archimedean copulas , 2011, Comput. Stat. Data Anal..

[28]  Tandy Warnow,et al.  SuperFine: fast and accurate supertree estimation. , 2012, Systematic biology.