Vine copula

Last updated

A vine is a graphical tool for labeling constraints in high-dimensional probability distributions. A regular vine is a special case for which all constraints are two-dimensional or conditional two-dimensional. Regular vines generalize trees, and are themselves specializations of Cantor tree. [1]

Contents

Combined with bivariate copulas, regular vines have proven to be a flexible tool in high-dimensional dependence modeling. Copulas [2] [3] are multivariate distributions with uniform univariate margins. Representing a joint distribution as univariate margins plus copulas allows the separation of the problems of estimating univariate distributions from the problems of estimating dependence. This is handy in as much as univariate distributions in many cases can be adequately estimated from data, whereas dependence information is roughly unknown, involving summary indicators and judgment. [4] [5] Although the number of parametric multivariate copula families with flexible dependence is limited, there are many parametric families of bivariate copulas. Regular vines owe their increasing popularity to the fact that they leverage from bivariate copulas and enable extensions to arbitrary dimensions. Sampling theory and estimation theory for regular vines are well developed [6] [7] and model inference has left the post . [8] [9] [7] Regular vines have proven useful in other problems such as (constrained) sampling of correlation matrices, [10] [11] building non-parametric continuous Bayesian networks. [12] [13]

For example, in finance, vine copulas have been shown to effectively model tail risk in portfolio optimization applications. [14]

Historical origins

The first regular vine, avant la lettre, was introduced by Harry Joe. [15] The motive was to extend parametric bivariate extreme value copula families to higher dimensions. To this end he introduced what would later be called the D-vine. Joe [16] was interested in a class of n-variate distributions with given one dimensional margins, and n(n − 1) dependence parameters, whereby n − 1 parameters correspond to bivariate margins, and the others correspond to conditional bivariate margins. In the case of multivariate normal distributions, the parameters would be n − 1 correlations and (n − 1)(n − 2)/2 partial correlations, which were noted to be algebraically independent in (−1, 1).

An entirely different motivation underlay the first formal definition of vines in Cooke. [17] Uncertainty analyses of large risk models, such as those undertaken for the European Union and the US Nuclear Regulatory Commission for accidents at nuclear power plants, involve quantifying and propagating uncertainty over hundreds of variables. [18] [19] [20] Dependence information for such studies had been captured with Markov trees, [21] which are trees constructed with nodes as univariate random variables and edges as bivariate copulas. For n variables, there are at most n − 1 edges for which dependence can be specified. New techniques at that time involved obtaining uncertainty distributions on modeling parameters by eliciting experts' uncertainties on other variables which are predicted by the models. These uncertainty distributions are pulled back onto the model's parameters by a process known as probabilistic inversion. [8] [18] The resulting distributions often displayed a dependence structure that could not be captured as a Markov tree.

Graphical models called vines were introduced in 1997 and further refined by Roger M. Cooke, Tim Bedford, and Dorota Kurowicka. [17] [1] [8] An important feature of vines is that they can add conditional dependencies among variables on top of a Markov tree which is generally too parsimonious to summarize the dependence among variables.

Regular vines (R-vines)

C-vine on 4 variables C-Vine on 4 variables.png
C-vine on 4 variables
D-vine on 4 variables D-Vine on 4 variables.png
D-vine on 4 variables
R-vine on 5 variables Regular vine on 5 variables.png
R-vine on 5 variables

A vine V on n variables is a nested set of connected trees where the edges in the first tree are the nodes of the second tree, the edges of the second tree are the nodes of the third tree, etc. A regular vine or R-vine on n variables is a vine in which two edges in tree j are joined by an edge in tree j + 1 only if these edges share a common node, j = 1, ..., n− 2. The nodes in the first tree are univariate random variables. The edges are constraints or conditional constraints explained as follows.

Recall that an edge in a tree is an unordered set of two nodes. Each edge in a vine is associated with a constraint set, being the set of variables (nodes in first tree) reachable by the set membership relation. For each edge, the constraint set is the union of the constraint sets of the edge's two members called its component constraint sets (for an edge in the first tree, the component constraint sets are empty). The constraint associated with each edge is now the symmetric difference of its component constraint sets conditional on the intersection of its constraint sets. One can show that for a regular vine, the symmetric difference of the component constraint sets is always a doubleton and that each pair of variables occurs exactly once as constrained variables. In other words, all constraints are bivariate or conditional bivariate.

The degree of a node is the number of edges attaching to it. The simplest regular vines have the simplest degree structure; the D-Vine assigns every node degree 1 or 2, the C-Vine assigns one node in each tree the maximal degree. For large vines, it is clearer to draw each tree separately.

The number of regular vines on n variables grows rapidly in n: there are 2n−3 ways of extending a regular vine with one additional variable, and there are n(n− 1)(n− 2)!2(n− 2)(n− 3)/2/2 labeled regular vines on n variables [22] . [23]

The constraints on a regular vine may be associated with partial correlations or with conditional bivariate copula. In the former case, we speak of a partial correlation vine, and in the latter case of a vine copula.

Partial correlation vines

Bedford and Cooke [1] show that any assignment of values in the open interval (−1, 1) to the edges in any partial correlation vine is consistent, the assignments are algebraically independent, and there is a one-to-one relation between all such assignments and the set of correlation matrices. In other words, partial correlation vines provide an algebraically independent parametrization of the set of correlation matrices, whose terms have an intuitive interpretation. Moreover, the determinant of the correlation matrix is the product over the edges of (1 −ρ2ik;D(ik)) where ρik;D(ik) is the partial correlation assigned to the edge with conditioned variables i,k and conditioning variables D(ik). A similar decomposition characterizes the mutual information, which generalizes the determinant of the correlation matrix. [17] These features have been used in constrained sampling of correlation matrices, [10] building non-parametric continuous Bayesian networks [12] [13] and addressing the problem of extending partially specified matrices to positive definite matrices [24] . [25]

Vine copulas or pair-copula construction

Under suitable differentiability conditions, any multivariate density f1...n on n variables, with univariate densities f1,...,fn, may be represented in closed form as a product of univariate densities and (conditional) copula densities on any R-vine V

[26]

f1...n = f1...fn Πe∈E(V) Ce1,e2|De ( Fe1|De , Fe2|De )

where edges e = (e1, e2) with conditioning set De are in the edge set E(V) of any regular vine V. The conditional copula densities Ce1,e2|De in this representation depend on the cumulative conditional distribution functions of the conditioned variables, Fe1|De , Fe2|De, and, potentially, on the values of the conditioning variables. When the conditional copulas do not depend on the values of the conditioning variables, one speaks of the simplifying assumption of constant conditional copulas. Though most applications invoke this assumption, exploring the modelling freedom gained by discharging this assumption has begun [27] [28] . [29] When bivariate Gaussian copulas are assigned to edges of a vine, then the resulting multivariate density is the Gaussian density parametrized by a partial correlation vine rather than by a correlation matrix.

The vine pair-copula construction, based on the sequential mixing of conditional distributions has been adapted to discrete variables and mixed discrete/continuous response [30] . [31] Also factor copulas, where latent variables have been added to the vine, have been proposed (e.g., [32] ).

Vine researchers have developed algorithms for maximum likelihood estimation and simulation of vine copulas, finding truncated vines that summarize the dependence in the data, enumerating through vines, etc. Chapter 6 of Dependence Modeling with Copulas [33] summarizes these algorithms in pseudocode.


Truncated vine copulas (introduced by E.C Brechmann in his Ph.D. thesis) are vine copulas that have independence copulas in the last trees. This way truncated vine copulas encode in their structure conditional independences. Truncated vines are very useful because they contain much fewer parameters than regular vines. An important question is what should be the tree at the highest level. An interesting relationship between truncated vines and cherry tree copulas is presented in ( [34] ) Cherry tree graph representations were introduced as an alternative for the usual graphical representations of vine copulas, moreover, the conditional independencies encoded by the last tree (first tree after truncation) is also highlighted here ( [35] ) and in ( [36] ) The cherry tree sequence representation of the vine copulas gives a new way to look at truncated copulas, based on the conditional independence which is caused by truncation.

Parameter estimation

For parametric vine copulas, with a bivariate copula family on each edge of a vine, algorithms and software are available for maximum likelihood estimation of copula parameters, assuming data have been transformed to uniform scores after fitting univariate margins. There are also available algorithms (e.g., [37] ) for choosing good truncated regular vines where edges of high-level trees are taken as conditional independence. These algorithms assign variables with strong dependence or strong conditional dependence to low order trees in order that higher order trees have weak conditional dependence or conditional independence. Hence parsimonious truncated vines are obtained for a large number of variables. Software with a user interface in R are available (e.g., [38] ).

Sampling and conditionalizing

A sampling order for n variables is a sequence of conditional densities in which the first density is unconditional, and the densities for other variables are conditioned on the preceding variables in the ordering. A sampling order is implied by a regular-vine representation of the density if each conditional density can be written as a product of copula densities in the vine and one dimensional margins. [23]

An implied sampling order is generated by a nested sequence of subvines where each sub-vine in the sequence contains one new variable not present in the preceding sub-vine. For any regular vine on n variables there are 2n−1 implied sampling orders. Implied sampling orders are a small subset of all n! orders but they greatly facilitate sampling. Conditionalizing a regular vine on values of an arbitrary subset of variables is a complex operation. However, conditionalizing on an initial sequence of an implied sampling order is trivial, one simply plugs in the initial conditional values and proceeds with the sampling. A general theory of conditionalization does not exist at present.

Further reading

Related Research Articles

Multivariate statistics is a subdivision of statistics encompassing the simultaneous observation and analysis of more than one outcome variable, i.e., multivariate random variables. Multivariate statistics concerns understanding the different aims and background of each of the different forms of multivariate analysis, and how they relate to each other. The practical application of multivariate statistics to a particular problem may involve several types of univariate and multivariate analyses in order to understand the relationships between variables and their relevance to the problem being studied.

<span class="mw-page-title-main">Multivariate normal distribution</span> Generalization of the one-dimensional normal distribution to higher dimensions

In probability theory and statistics, the multivariate normal distribution, multivariate Gaussian distribution, or joint normal distribution is a generalization of the one-dimensional (univariate) normal distribution to higher dimensions. One definition is that a random vector is said to be k-variate normally distributed if every linear combination of its k components has a univariate normal distribution. Its importance derives mainly from the multivariate central limit theorem. The multivariate normal distribution is often used to describe, at least approximately, any set of (possibly) correlated real-valued random variables each of which clusters around a mean value.

<span class="mw-page-title-main">Extreme value theory</span> Branch of statistics focusing on large deviations

Extreme value theory or extreme value analysis (EVA) is a branch of statistics dealing with the extreme deviations from the median of probability distributions. It seeks to assess, from a given ordered sample of a given random variable, the probability of events that are more extreme than any previously observed. Extreme value analysis is widely used in many disciplines, such as structural engineering, finance, economics, earth sciences, traffic prediction, and geological engineering. For example, EVA might be used in the field of hydrology to estimate the probability of an unusually large flooding event, such as the 100-year flood. Similarly, for the design of a breakwater, a coastal engineer would seek to estimate the 50-year wave and design the structure accordingly.

<span class="mw-page-title-main">Correlation</span> Statistical concept

In statistics, correlation or dependence is any statistical relationship, whether causal or not, between two random variables or bivariate data. Although in the broadest sense, "correlation" may indicate any type of association, in statistics it usually refers to the degree to which a pair of variables are linearly related. Familiar examples of dependent phenomena include the correlation between the height of parents and their offspring, and the correlation between the price of a good and the quantity the consumers are willing to purchase, as it is depicted in the so-called demand curve.

<span class="mw-page-title-main">Spearman's rank correlation coefficient</span> Nonparametric measure of rank correlation

In statistics, Spearman's rank correlation coefficient or Spearman's ρ, named after Charles Spearman and often denoted by the Greek letter (rho) or as , is a nonparametric measure of rank correlation. It assesses how well the relationship between two variables can be described using a monotonic function.

<span class="mw-page-title-main">Multivariate analysis of variance</span> Procedure for comparing multivariate sample means

In statistics, multivariate analysis of variance (MANOVA) is a procedure for comparing multivariate sample means. As a multivariate procedure, it is used when there are two or more dependent variables, and is often followed by significance tests involving individual dependent variables separately.

The general linear model or general multivariate regression model is a compact way of simultaneously writing several multiple linear regression models. In that sense it is not a separate statistical linear model. The various multiple linear regression models may be compactly written as

In probability theory and statistics, a copula is a multivariate cumulative distribution function for which the marginal probability distribution of each variable is uniform on the interval [0, 1]. Copulas are used to describe/model the dependence (inter-correlation) between random variables. Their name, introduced by applied mathematician Abe Sklar in 1959, comes from the Latin for "link" or "tie", similar but unrelated to grammatical copulas in linguistics. Copulas have been used widely in quantitative finance to model and minimize tail risk and portfolio-optimization applications.

<span class="mw-page-title-main">Estimation of distribution algorithm</span>

Estimation of distribution algorithms (EDAs), sometimes called probabilistic model-building genetic algorithms (PMBGAs), are stochastic optimization methods that guide the search for the optimum by building and sampling explicit probabilistic models of promising candidate solutions. Optimization is viewed as a series of incremental updates of a probabilistic model, starting with the model encoding an uninformative prior over admissible solutions and ending with the model that generates only the global optima.

In statistics, the multivariate t-distribution is a multivariate probability distribution. It is a generalization to random vectors of the Student's t-distribution, which is a distribution applicable to univariate random variables. While the case of a random matrix could be treated within this structure, the matrix t-distribution is distinct and makes particular use of the matrix structure.

In probability theory, comonotonicity mainly refers to the perfect positive dependence between the components of a random vector, essentially saying that they can be represented as increasing functions of a single random variable. In two dimensions it is also possible to consider perfect negative dependence, which is called countermonotonicity.

Portfolio optimization is the process of selecting the best portfolio, out of the set of all portfolios being considered, according to some objective. The objective typically maximizes factors such as expected return, and minimizes costs like financial risk. Factors being considered may range from tangible to intangible.

In probability and statistics, an elliptical distribution is any member of a broad family of probability distributions that generalize the multivariate normal distribution. Intuitively, in the simplified two and three dimensional case, the joint distribution forms an ellipse and an ellipsoid, respectively, in iso-density plots.

Financial correlations measure the relationship between the changes of two or more financial variables over time. For example, the prices of equity stocks and fixed interest bonds often move in opposite directions: when investors sell stocks, they often use the proceeds to buy bonds and vice versa. In this case, stock and bond prices are negatively correlated.

Mir Maswood Ali was a Canadian statistician and mathematician of Bengali origin. He is known for co-discovering the Ali-Mikhail-Haq copula, which is a topic of active research, both in theory and application. Ali played a key role in establishing the Journal of Statistical Research, of which the first issue appeared in 1970. The December 2008 issue of the Journal of Statistical Research was dedicated in honor of Ali. In 2008, Ali received the Qazi Motahar Husain Gold Medal Award in recognition of his contributions to statistics.

In the mathematical theory of probability, multivariate Laplace distributions are extensions of the Laplace distribution and the asymmetric Laplace distribution to multiple variables. The marginal distributions of symmetric multivariate Laplace distribution variables are Laplace distributions. The marginal distributions of asymmetric multivariate Laplace distribution variables are asymmetric Laplace distributions.

A copula is a mathematical function that provides a relationship between marginal distributions of random variables and their joint distributions. Copulas are important because it represents a dependence structure without using marginal distributions. Copulas have been widely used in the field of finance, but their use in signal processing is relatively new. Copulas have been employed in the field of wireless communication for classifying radar signals, change detection in remote sensing applications, and EEG signal processing in medicine. In this article, a short introduction to copulas is presented, followed by a mathematical derivation to obtain copula density functions, and then a section with a list of copula density functions with applications in signal processing.

In probability theory and Bayesian statistics, the Lewandowski-Kurowicka-Joe distribution, often referred to as the LKJ distribution, is a probability distribution over positive definite symmetric matrices with unit diagonals. It is commonly used as a prior for correlation matrix in hierarchical Bayesian modeling. Hierarchical Bayesian modeling often tries to make an inference on the covariance structure of the data, which can be decomposed into a scale vector and correlation matrix. Instead of the prior on the covariance matrix such as the inverse-Wishart distribution, LKJ distribution can serve as a prior on the correlation matrix along with some suitable prior distribution on the scale vector. The distribution was first introduced in a more general context and is an example of the vine copula, an approach to constrained high-dimensional probability distributions. It has been implemented as part of the Stan probabilistic programming language and as a library linked to the Turing.jl probabilistic programming library in Julia.

<span class="mw-page-title-main">Homoscedasticity and heteroscedasticity</span> Statistical property

In statistics, a sequence of random variables is homoscedastic if all its random variables have the same finite variance; this is also known as homogeneity of variance. The complementary notion is called heteroscedasticity, also known as heterogeneity of variance. The spellings homoskedasticity and heteroskedasticity are also frequently used. Assuming a variable is homoscedastic when in reality it is heteroscedastic results in unbiased but inefficient point estimates and in biased estimates of standard errors, and may result in overestimating the goodness of fit as measured by the Pearson coefficient.

References

  1. 1 2 3 Bedford, T.J.; Cooke, R.M. (2002). "Vines — a new graphical model for dependent random variables". Annals of Statistics. 30 (4): 1031–1068. CiteSeerX   10.1.1.26.8965 . doi:10.1214/aos/1031689016.
  2. ↑ Joe, H. (1997). Multivariate Models and Dependence Concepts. London: Chapman & Hall.
  3. ↑ Nelsen, R.B. (2006). An Introduction to Copulas, 2nd ed. New York: Springer.
  4. ↑ Kraan, B.C.P.; Cooke, R.M. (2000). "Processing expert judgements in accident consequence modeling". Radiation Protection Dosimetry. 90 (3): 311–315. doi:10.1093/oxfordjournals.rpd.a033153.
  5. ↑ Ale, B.J.M.; Bellamy, L.J.; van der Boom, R.; Cooper, J.; Cooke, R.M.; Goossens, L.H.J.; Hale, A.R.; Kurowicka, D.; Morales, O.; Roelen, A.L.C.; Spouge, J. (2009). "Further development of a Causal model for Air Transport Safety (CATS): Building the mathematical heart". Reliability Engineering and System Safety Journal. 94 (9): 1433–1441. doi:10.1016/j.ress.2009.02.024.
  6. ↑ Kurowicka, D.; Cooke, R.M. (2007). "Sampling algorithms for generating joint uniform distributions using the vine-copula method". Computational Statistics and Data Analysis. 51 (6): 2889–2906. doi:10.1016/j.csda.2006.11.043.
  7. 1 2 Aas, K.; Czado, C.; Frigessi, A.; Bakken, H. (2009). "Pair-copula constructions of multiple dependence". Insurance: Mathematics and Economics. 44 (2): 182–198. CiteSeerX   10.1.1.61.3984 . doi:10.1016/j.insmatheco.2007.02.001. S2CID   18320750.
  8. 1 2 3 Kurowicka, D.; Cooke, R.M. (2006). Uncertainty Analysis with High Dimensional Dependence Modelling. Wiley.
  9. ↑ Kurowicka, D.; Cooke, R.M.; Callies, U. (2007). "Vines inference". Brazilian Journal of Probability and Statistics.
  10. 1 2 Lewandowski, D.; Kurowicka, D.; Joe, H. (2009). "Generating random correlation matrices based on vines and extended onion method". Journal of Multivariate Analysis. 100 (9): 1989–2001. doi: 10.1016/j.jmva.2009.04.008 .
  11. ↑ Kurowicka, D. (2014). "Generating random correlation matrices based on vines and extended onion method". Joint Density of Correlations in the Correlation Matrix with Chordal Sparsity Patterns. 129 (C): 160–170. doi: 10.1016/j.jmva.2014.04.006 .
  12. 1 2 Hanea, A.M. (2008). Algorithms for Non-parametric Bayesian Belief Nets (Ph.D.). Delft Institute of Applied Mathematics, Delft University of Technology.
  13. 1 2 Hanea, A.M.; Kurowicka, D.; Cooke, R.M.; Ababei, D.A. (2010). "Mining and visualising ordinal data with non-parametric continuous BBNs". Computational Statistics and Data Analysis. 54 (3): 668–687. doi:10.1016/j.csda.2008.09.032.
  14. ↑ Low, R.K.Y.; Alcock, J.; Faff, R.; Brailsford, T. (2013). "Canonical vine copulas in the context of modern portfolio management: Are they worth it?". Journal of Banking & Finance. 37 (8): 3085–3099. doi:10.1016/j.jbankfin.2013.02.036. S2CID   154138333.
  15. ↑ Joe, H. (1994). "Multivariate extreme-value distributions with applications in environmental data". The Canadian Journal of Statistics. 22 (1): 47–64. doi:10.2307/3315822. JSTOR   3315822.
  16. ↑ Joe, H. (1996), "Families of m-variate distributions with given margins and m(m−1)/2 bivariate dependence parameters", in Rüschendorf, L.; Schweizer, B.; Taylor, M.D. (eds.), Distributions with fixed marginals and related topics, vol. 28, pp. 120–141
  17. 1 2 3 Cooke, R.M. (1997). "Markov and entropy properties of tree and vine dependent variables". Proc. ASA Section of Bayesian Statistical Science.
  18. 1 2 Goossens, L.H.J.; Harper, F.T.; Kraan, B.C.P.; Metivier, H. (2000). "Expert judgement for a probabilistic accident consequence uncertainty analysis". Radiation Protection Dosimetry. 90 (3): 295–301. doi:10.1093/oxfordjournals.rpd.a033151.
  19. ↑ Harper, F.; Goossens, L.H.J.; Cooke, R.M.; Hora, S.; Young, M.; Pasler-Ssauer, J.; Miller, L.; Kraan, B.C.P.; Lui, C.; McKay, M.; Helton, J.; Jones, A. (1994), Joint USNRC CEC consequence uncertainty study: Summary of objectives, approach, application, and results for the dispersion and deposition uncertainty assessment, vol. III, NUREG/CR-6244, EUR 15755 EN, SAND94-1453
  20. ↑ Guégan, D.; Hassani, B.K. (2013), "Multivariate VaRs for operational risk capital computation: a vine structure approach", International Journal of Risk Assessment and Management, 17 (2): 148–170, CiteSeerX   10.1.1.686.4277 , doi:10.1504/IJRAM.2013.057104, S2CID   4989901
  21. ↑ Whittaker, J. (1990). Graphical Models in Applied Multivariate Statistics. Chichester: Wiley.
  22. ↑ Morales Napoles, O.; Cooke, R.M.; Kurowicka, D. (2008), The number of vines and regular vines on n nodes, vol. Technical report, Delft Institute of Applied Mathematics, Delft University of Technology
  23. 1 2 Cooke, R.M.; Kurowicka, D.; Wilson, K. (2015). "Sampling, conditionalizing, counting, merging, searching regular vines". Journal of Multivariate Analysis. 138: 4–18. doi: 10.1016/j.jmva.2015.02.001 .
  24. ↑ Kurowicka, D.; Cooke, R.M. (2003). "A parametrization of positive definite matrices in terms of partial correlation vines". Linear Algebra and Its Applications. 372: 225–251. doi: 10.1016/s0024-3795(03)00507-x .
  25. ↑ Kurowicka, D.; Cooke, R.M. (2006). "Completion problem with partial correlation vines". Linear Algebra and Its Applications. 418 (1): 188–200. doi:10.1016/j.laa.2006.01.031.
  26. ↑ Beford, T.J.; Cooke, R.M. (2001). "Probability density decomposition for conditionally dependent random variables modeled by vines". Annals of Mathematics and Artificial Intelligence. 32: 245–268. doi:10.1023/A:1016725902970. S2CID   42550420.
  27. ↑ Hobaek Haff, I.; Aas, K.; Frigessi, A. (2010). "On the simplified pair-copula construction - simply useful or too simplistic?". Journal of Multivariate Analysis. 101 (5): 1296–1310. doi:10.1016/j.jmva.2009.12.001. hdl: 10852/34736 .
  28. ↑ Acar, E.F.; Genest, C.; Nešlehová, J. (2012). "Beyond simplified pair-copula constructions". Journal of Multivariate Analysis. 110: 74–90. doi: 10.1016/j.jmva.2012.02.001 .
  29. ↑ Stoeber, J.; Joe, H.; Czado, C. (2013). "Simplified pair copula constructions, limitations and extensions". Journal of Multivariate Analysis. 119: 101–118. doi: 10.1016/j.jmva.2013.04.014 .
  30. ↑ Panagiotelis, A.; Czado, C.; Joe, H. (2012). "Regular vine distributions for discrete data". Journal of the American Statistical Association. 105 (499): 1063–1072. doi:10.1080/01621459.2012.682850. S2CID   123502012.
  31. ↑ Stoeber, J.; Hong, H.G.; Czado, C.; Ghosh, P. (2015). "Comorbidity of chronic diseases in the elderly: Patterns identified by a copula design for mixed responses". Computational Statistics and Data Analysis. 88: 28–39. doi:10.1016/j.csda.2015.02.001.
  32. ↑ Krupskii, P.; Joe, H. (2013). "Factor copula models for multivariate data". Journal of Multivariate Analysis. 120: 85–101. doi:10.1016/j.jmva.2013.05.001.
  33. ↑ Joe, H. (2014). Dependence Modeling with Copulas. Chapman Hall. ISBN   978-1-4665-8322-1.
  34. ↑ Kovacs, E.; Szantai, T. (2017). "On the connection between cherry-tree copulas and truncated R-vine copulas". Kybernetica. 53 (3): 437–460. arXiv: 1604.03269 . doi:10.14736/kyb-2017-3-0437. S2CID   45343495.
  35. ↑ Kovacs, E.; Szantai, T. (2012), "Vine copulas as a mean for the construction of high dimensional probability distribution associated to a Markov Network", arXiv: 1105.1697 [math.ST]
  36. ↑ Kovacs, E.; Szantai, T. (2012). "Hypergraphs in the characterization of regular-vine copula structures". Proc. 13th International Conference on Mathematics and Its Applications, Timisoara. 2012(a): 335–344.
  37. ↑ Brechmann, E.C.; Czado, C.; Aas, K. (2012). "Truncated regular vines in high dimensions with application to financial data". Canadian Journal of Statistics. 40 (1): 68–85. CiteSeerX   10.1.1.185.2933 . doi:10.1002/cjs.10141. S2CID   2155236.
  38. ↑ Schepsmeier, U.; Stoeber, J.; Brechmann, E.C.; Graeler, B. (2014). "Vine Copula: Statistical inference of vine copulas, R package version 1.3".
This page is based on this Wikipedia article
Text is available under the CC BY-SA 4.0 license; additional terms may apply.
Images, videos and audio are available under their respective licenses.