Abstract
Discrete vine copulas provide a flexible modeling framework for high-dimensional data and have significant computational advantages over competing methods. A vine-based multivariate probability mass function is constructed from bivariate copula building blocks and univariate marginal distributions. However, even for a moderate number of variables, the number of alternative vine decompositions is very large and additionally there is a large set of candidate bivariate copula families that can be used as building blocks in any given decomposition. Together, these two issues ensure that it is infeasible to evaluate all possible vine copula models. Instead, two greedy algorithms for automatically selecting vine structures and component pair-copula building blocks are introduced. The algorithms are tested in a simulation study that is itself driven by real world data from online retail. Both algorithms select vines that provide accurate estimates of the joint probabilities. Using three different f-divergences as criteria, the proposed algorithms outperform a Gaussian copula benchmark, especially for data with high dependence. Finally, the selection algorithm is applied to data from the General Social Survey and outperforms a Gaussian copula benchmark using both in-sample and out-of-sample criteria.
| Original language | English |
|---|---|
| Pages (from-to) | 138-152 |
| Number of pages | 15 |
| Journal | Computational Statistics and Data Analysis |
| Volume | 106 |
| DOIs | |
| Publication status | Published - 1 Feb 2017 |
Keywords
- Count data
- Model selection
- Tail asymmetry
- Tail dependence
Cite this
}
Model selection for discrete regular vine copulas. / Panagiotelis, Anastasios; Czado, Claudia; Joe, Harry; Stöber, Jakob.
In: Computational Statistics and Data Analysis, Vol. 106, 01.02.2017, p. 138-152.Research output: Contribution to journal › Article › Research › peer-review
TY - JOUR
T1 - Model selection for discrete regular vine copulas
AU - Panagiotelis, Anastasios
AU - Czado, Claudia
AU - Joe, Harry
AU - Stöber, Jakob
PY - 2017/2/1
Y1 - 2017/2/1
N2 - Discrete vine copulas provide a flexible modeling framework for high-dimensional data and have significant computational advantages over competing methods. A vine-based multivariate probability mass function is constructed from bivariate copula building blocks and univariate marginal distributions. However, even for a moderate number of variables, the number of alternative vine decompositions is very large and additionally there is a large set of candidate bivariate copula families that can be used as building blocks in any given decomposition. Together, these two issues ensure that it is infeasible to evaluate all possible vine copula models. Instead, two greedy algorithms for automatically selecting vine structures and component pair-copula building blocks are introduced. The algorithms are tested in a simulation study that is itself driven by real world data from online retail. Both algorithms select vines that provide accurate estimates of the joint probabilities. Using three different f-divergences as criteria, the proposed algorithms outperform a Gaussian copula benchmark, especially for data with high dependence. Finally, the selection algorithm is applied to data from the General Social Survey and outperforms a Gaussian copula benchmark using both in-sample and out-of-sample criteria.
AB - Discrete vine copulas provide a flexible modeling framework for high-dimensional data and have significant computational advantages over competing methods. A vine-based multivariate probability mass function is constructed from bivariate copula building blocks and univariate marginal distributions. However, even for a moderate number of variables, the number of alternative vine decompositions is very large and additionally there is a large set of candidate bivariate copula families that can be used as building blocks in any given decomposition. Together, these two issues ensure that it is infeasible to evaluate all possible vine copula models. Instead, two greedy algorithms for automatically selecting vine structures and component pair-copula building blocks are introduced. The algorithms are tested in a simulation study that is itself driven by real world data from online retail. Both algorithms select vines that provide accurate estimates of the joint probabilities. Using three different f-divergences as criteria, the proposed algorithms outperform a Gaussian copula benchmark, especially for data with high dependence. Finally, the selection algorithm is applied to data from the General Social Survey and outperforms a Gaussian copula benchmark using both in-sample and out-of-sample criteria.
KW - Count data
KW - Model selection
KW - Tail asymmetry
KW - Tail dependence
UR - http://www.scopus.com/inward/record.url?scp=84992469998&partnerID=8YFLogxK
U2 - 10.1016/j.csda.2016.09.007
DO - 10.1016/j.csda.2016.09.007
M3 - Article
VL - 106
SP - 138
EP - 152
JO - Computational Statistics and Data Analysis
JF - Computational Statistics and Data Analysis
SN - 0167-9473
ER -