Abstract
Vine copulas (or pair-copula constructions) have become an important tool for high-dimensional dependence modeling. Typically, so-called simplified vine copula models are estimated where bivariate conditional copulas are approximated by bivariate unconditional copulas. We present the first nonparametric estimator of a non-simplified vine copula that allows for varying conditional copulas using penalized hierarchical B-splines. Throughout the vine copula, we test for the simplifying assumption in each edge, establishing a data-driven non-simplified vine copula estimator. To overcome the curse of dimensionality, we approximate conditional copulas with more than one conditioning argument by a conditional copula with the first principal component as conditioning argument. An extensive simulation study is conducted, showing a substantial improvement in the out-of-sample Kullback–Leibler divergence if the null hypothesis of a simplified vine copula can be rejected. We apply our method to the famous uranium data and present a classification of an eye state data set, demonstrating the potential benefit that can be achieved when conditional copulas are modeled.
Similar content being viewed by others
Notes
We do not use the asymptotic normal distribution of Vuong’s test because both over-rejection and under-rejection can be severe when the competing models contain a rather large number of parameters, see Shi (2015). Moreover, we use an out-of-sample version of Vuong’s test because the in-sample version does not account for the number of parameters.
Note that \(\beta \) could be increased up to the value of one, so that the rather strong variation for \(\beta =0.6\) is by far not the strongest possible variation.
We also present tables with all results in the supplementary material. Table 1 in the supplementary material shows the simulation results for the three-dimensional non-simplified vine copulas for the scenario case (a), while Table 2 in the supplementary material presents the results for case (b), where the lowest mean of the out-of-sample KL divergence measure for the nonparametric estimations “Test” is set in bold.
Table 5 in the supplementary material contains the simulation results for the three-dimensional normal distribution and five-dimensional normal distribution, where the lowest mean of the out-of-sample KL divergence measure for the nonparametric estimations “Test” is set in bold.
Table 3 in the supplementary material shows the simulation results for the five-dimensional non-simplified vine copulas for the scenario case (a), while Table 4 in the supplementary material presents the results for case (b), where the lowest mean of the out-of-sample KL divergence measure for the nonparametric estimations “Test” is set in bold.
For \(l=2\), we get \(\mathcal{I}_2 = \{2,4\}, \tau (2)=(0,0.25,0.5,0.75,1)\), and
$$\begin{aligned} {B}^{(\tau (2))}_{\mathcal {I}_2}(\mathbf {u}_j)&= B^{(\tau (2))}(\mathbf {u}_j) \begin{pmatrix}0 &{} 0 &{} 0 &{} 1 &{} 0 \\ 0 &{} 1 &{} 0 &{} 0 &{} 0 \end{pmatrix}^\top \\&= \begin{pmatrix} \phi _{0.25}(u_{j,1}) &{} \phi _{0.75}(u_{j,1})\\ \vdots &{} \vdots \\ \phi _{0.25}(u_{j,n}) &{} \phi _{0.75}(u_{j,n}) \end{pmatrix}. \end{aligned}$$
References
Aas, K., Berg, D.: Models for construction of multivariate dependence: a comparison study. Eur. J. Finance 15(7–8), 639–659 (2009)
Aas, K., Czado, C., Frigessi, A., Bakken, H.: Pair-copula constructions of multiple dependence. Insur. Math. Econ. 44(2), 182–198 (2009)
Abegaz, F., Gijbels, I., Veraverbeke, N.: Semiparametric estimation of conditional copulas. J. Multivar. Anal. 110, 43–73 (2012). Special Issue on Copula Modeling and Dependence
Acar, E.F., Craiu, R.V., Yao, F.: Dependence calibration in conditional copulas: a nonparametric approach. Biometrics 67(2), 445–453 (2011)
Acar, E.F., Genest, C., Nešlehová, J.: Beyond simplified pair-copula constructions. J. Multivar. Anal. 110, 74–90 (2012)
Bedford, T., Cooke, R.M.: Vines: a new graphical model for dependent random variables. Ann. Stat. 30(4), 1031–1068 (2002)
Brechmann, E.C., Czado, C., Aas, K.: Truncated regular vines in high dimensions with application to financial data. Can. J Stat. 40(1), 68–85 (2012)
Dennis Cook, R., Johnson, M.E.: Generalized burr-pareto-logistic distributions with applications to a uranium exploration data set. Technometrics 28(2), 123–131 (1986)
Dißmann, J., Brechmann, E., Czado, C., Kurowicka, D.: Selecting and estimating regular vine copulae and application to financial returns. Comput. Statist. Data Anal. 59, 52–69 (2013)
Duong, T.: ks: Kernel Smoothing. R package version 1.10.4 (2016)
Efron, B.: Selection criteria for scatterplot smoothers. Ann. Statist. 29, 470–504 (2001)
Eilers, P.H.C., Marx, B.D.: Flexible smoothing with B-splines and penalties. Statist. Sci. 11(2), 89–121 (1996)
Fermanian, J.-D., Wegkamp, M.H.: Time-dependent copulas. J. Multivar. Anal. 110, 19–29 (2012). Special Issue on Copula Modeling and Dependence
Fischer, M., Köck, C., Schlüter, S., Weigert, F.: An empirical analysis of multivariate copula models. Quant. Finance 9(7), 839–854 (2009)
Forsey, D.R., Bartels, R.H.: Hierarchical B-spline refinement. In SIGGRAPH ’88: Proceedings of the 15th annual conference on Computer graphics and interactive techniques, New York, NY, pp. 205–212. ACM (1988)
Forsey, D.R., Bartels, R.H.: Surface fitting with hierarchical splines. ACM Trans. Graph. 14(2), 134–161 (1995)
Gijbels, I., Omelka, M., Veraverbeke, N.: Multivariate and functional covariates and conditional copulas. Electron. J. Stat. 6, 1273–1306 (2012)
Gijbels, I., Veraverbeke, N., Omelka, M.: Conditional copulas, association measures and their applications. Comput. Stat. Data Anal. 55(5), 1919–1932 (2011)
Hall, P., Yao, Q.: Approximating conditional distribution functions using dimension reduction. Ann. Statist. 33(3), 1404–1421 (2005, 06)
Hobæk Haff, I., Aas, K., Frigessi, A.: On the simplified pair-copula construction: simply useful or too simplistic? J. Multivar. Anal. 101(5), 1296–1310 (2010)
Hurvich, C.M., Tsai, C.-L.: Regression and time series model selection in small samples. Biometrika 76(2), 297–307 (1989)
Joe, H.: Dependence modeling with copulas. CRC Press, Boca Raton (2015)
Kauermann, G., Schellhase, C.: Flexible pair-copula estimation in D-vines with penalized splines. Stat. Comput. 24(6), 1081–1100 (2014)
Kauermann, G., Schellhase, C., Ruppert, D.: Flexible copula density estimation with penalized hierarchical B-splines. Scand. J. Stat. 40(4), 685–705 (2013)
Killiches, M., Kraus, D., Czado, C.: Examination and visualisation of the simplifying assumption for vine copulas in three dimensions. ArXiv e-prints (2016, February)
Kurowicka, D., Joe, H. (eds.): Dependence modeling. World Scientific, Singapore (2011)
Kurz, M.S.: pacotest: Testing for Partial Copulas and the Simplifying Assumption in Vine Copulas, R package version 0.2 (2017)
Kurz, M.S., Spanhel, F.: Testing the simplifying assumption in high-dimensional vine copulas. In preparation (2016)
Lopez-Paz, D., Hernandez-Lobato, J., Ghahramani, Z.: Gaussian process vine copulas for multivariate dependence. In Proceedings of the 30th International Conference on Machine Learning, W&CP 28(2), pp. 10–18. JMLR (2013)
Marx, B., Eilers, P.H.C.: Multidimensional penalized signal regression. Technometrics 47, 13–22 (2005)
Nagler, T., Czado, C.: Evading the curse of dimensionality in nonparametric density estimation with simplified vine copulas. J. Multivar. Anal. 151, 69–89 (2016)
Nelsen, R.: An introduction to copulas, 2nd edn. Springer, Berlin (2006)
Patton, A.J.: Modelling asymmetric exchange rate dependence. Int. Econ. Rev. 47(2), 527–556 (2006)
Rue, H., Martino, S., Chopin, N.: Approximate Bayesian inference for latent Gaussian models by using integrated nested Laplace approximations. J. Roy. Stat. Soc. B 71(2), 319–392 (2009)
Ruppert, D., Wand, M., Carroll, R.: Semiparametric Regression. Cambridge University Press, Cambridge (2003)
Schall, R.: Estimation in generalized linear models with random effects. Biometrika 78(4), 719–727 (1991)
Schepsmeier, U., Stoeber, J., Brechmann, E.C., Graeler, B., Nagler, T., Erhardt, T.: VineCopula: Statistical Inference of Vine Copulas. R package version 2.0.5 (2016)
Scott, D.W.: The Curse of Dimensionality and Dimension Reduction. Wiley, Hoboken (2008)
Shi, X.: A nondegenerate vuong test. Quant. Econ. 6(1), 85–121 (2015)
Sklar, A.: Fonctions de répartition à n dimensions et leurs marges. Publ. Inst. Statist. Univ. Paris 8, 229–231 (1959)
Spanhel, F., Kurz, M.S.: Simplified vine copula models: Approximations based on the simplifying assumption. ArXiv e-prints (2015)
Spanhel, F., Kurz, M.S.: The partial copula: properties and associated dependence measures. Stat. Probab. Lett. 119, 76–83 (2016)
Stein, M.L.: A comparison of generalized cross validation and modified maximum likelihood for estimating the parameters of a stochastic process. Ann. Stat. 18, 1139–1157 (1990)
Stöber, J., Joe, H., Czado, C.: Simplified pair copula constructions–Limitations and extensions. J. Multivar. Anal. 119, 101–118 (2013)
Vatter, T., Chavez-Demoulin, V.: Generalized additive models for conditional dependence structures. J. Multivar. Anal. 141, 147–167 (2015)
Veraverbeke, N., Omelka, M., Gijbels, I.: Estimation of a conditional copula and association measures. Scand. J. Stat. 38(4), 766–780 (2011)
Wahba, G.: A comparison of GCV and GML for choosing the smoothing parameter in the generalized spline smoothing problem. Ann. Stat. 13, 1378–1402 (1985)
Wand, M., Jones, M.C.: Kernel smoothing. Chapman and Hall, London (1995)
Wand, M.P., Ormerod, J.T.: On semiparametric regression with O’Sullivan penalised splines. Aust. N. Z. J. Stat. 50(2), 179–198 (2008)
Wood, S.: Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. J. Roy. Stat. Soc. B 73(1), 3–36 (2011)
Zenger, C.: Sparse grids. Notes Numer. Fluid Mech. Multidiscip. Des. 31, 241–251 (1991)
Acknowledgements
We thank Göran Kauermann (LMU Munich) for discussions at the beginning of this project.
Author information
Authors and Affiliations
Corresponding author
Electronic supplementary material
Below is the link to the electronic supplementary material.
Appendices
Appendix 1: Construction of sparse B-spline density basis
In the following, objects signed with superscript \(\tilde{\ }\) and \(^{(d)}\) are associated with hierarchical B-spline basis functions, whereas objects signed with superscript \(\tilde{\ }\) and \(^{(d,D)}\) are linked to sparse B-spline basis functions. In order to transform \(B^{(\tau (d))}(\mathbf {u}_j)\) in (6) into its hierarchical representation (see Forsey and Bartels 1988, 1995), we define the hierarchical index sets \(\mathcal {I}_0=\{1, 2\}\) and \(\mathcal {I}_l = \{2j\mid j\in \mathbb {N},1\le j \le 2^{l-1}\}\), \(l=1,\dots ,d\). Let \(B^{(\tau (l))}_{\mathcal {I}_l}(\mathbf {u}_j)\) denote the columns \(\mathcal {I}_l\) of \(B^{\tau (l)}(\mathbf {u}_j)\).Footnote 8
The univariate hierarchical B-spline basis of degree d is defined as
Figure 2 presents the univariate B-spline basis \(B^{(\tau (d))}(\mathbf {u}_j)\) and the building parts of the corresponding hierarchical basis \(\tilde{\mathbf {B}}^{(\tau (d))}(\mathbf {u}_j)\) for \(d=2\). Let \(\tilde{e}\) be a K-dimensional row vector such that its kth element is given by \(\tilde{e}_k:= \min \{l=0,\dots ,d: k\le |\tau (l)|\}\), e.g., \(\tilde{e}= (0,0,1,2,2)\) if \(d=2\). The vector \(\tilde{e}\) denotes the hierarchical level of \(\tilde{\mathbf {B}}^{(\tau (d))}(\mathbf {u}_j)\), and its kth element identifies the hierarchical level of the kth column of \(\tilde{\mathbf {B}}^{(\tau (d))}(\mathbf {u}_j)\). By construction, \(B^{(\tau (d))}(\mathbf {u}_j)\) and \(\tilde{\mathbf {B}}^{(\tau (d))}(\mathbf {u}_j)\) have full rank, i.e., \(B^{(\tau (d))}(\mathbf {u}_j)\tilde{A}=\tilde{\mathbf {B}}^{(\tau (d))}(\mathbf {u}_j)\) for some invertible \(K \times K\) matrix \(\tilde{A}\), so that both univariate bases \(\tilde{\mathbf {B}}^{(\tau (d))}(\mathbf {u}_j)\) and \(B^{(\tau (d))}(\mathbf {u}_j)\) span the same space. The three-dimensional hierarchical B-spline basis follows as
and the corresponding approximation of the conditional copula density is given by
where \(\tilde{\mathbf {b}}^{(d)} = (\bigotimes _{j=1}^3\tilde{A})^{-1}\mathbf {b} \).
To overcome the exponential increase in the number of spline coefficients, we use a three-dimensional sparse B-spline basis which reduces the dimension by deleting the columns from the full tensor product basis whose cumulated hierarchy level exceeds D, where \(d\le D\le 3d\). The cumulated hierarchy level of the full tensor product basis is defined as follows. For \(\alpha \in \mathbb {R}^{1\times n}\) and \(\beta \in \mathbb {R}^{1\times q}\), define the lth element of \((\alpha \oplus \beta )\in \mathbb {R}^{1\times nq}\) by \((\alpha \oplus \beta )_{l} = \alpha _{\left\lceil \frac{l}{q}\right\rceil }+ \beta _{l-q(\left\lceil \frac{l}{q}\right\rceil -1)}\), where \(\lceil \cdot \rceil \) is the ceil function. Note that the operation \(\oplus \) is associative. Recall that the kth element of \(\tilde{e}\) identifies the hierarchy level of the kth column of \(\tilde{\mathbf {B}}^{(\tau (d))}(\mathbf {u}_j)\), so that the lth element of \(\epsilon = \tilde{e}\oplus \tilde{e}\oplus \tilde{e}\in \mathbb {R}^{1\times {K}^3}\) denotes the cumulated hierarchy level of the lth column of the hierarchical B-spline basis \(\tilde{\varvec{\mathbf {\Phi }}}^{(d)}(\mathbf {u}_1,\mathbf {u}_2,\mathbf {u}_3)\). Define \(\mathcal{O}_D = \{j=1,\ldots , {K_{}}^3 :\epsilon _j\le D \}\), i.e., \(\mathcal{O}_D\) contains the position of the columns of \(\tilde{\varvec{\mathbf {\Phi }}}^{(d)}(\mathbf {u}_1,\mathbf {u}_2,\mathbf {u}_3)\) whose cumulated hierarchy level does not exceed D. Let \(\mathcal{O}_D(j)\) be the jth smallest element of \(\mathcal{O}_D\) and define the orthogonal matrix \(\mathcal {E}(\mathcal {O}_D)\in \mathbb {R}^{{K}^3\times |\mathcal{O}_D|}\) such that its \((\mathcal {O}_D(j),j)\)-entry is one for \(j=1,\dots ,|{\mathcal{O}}_D|,\) and the other entries are zero. The three-dimensional sparse B-spline basis follows as
where the lower index d is the degree of the univariate hierarchical B-spline basis and the upper index \(D, d\le D\le 3d,\) refers to the maximum cumulated hierarchy level. Only the columns in the hierarchical B-spline basis \({\tilde{\varvec{\Phi }}}^{(d)}(\mathbf {u}_1,\mathbf {u}_2,\mathbf {u}_3)\) whose cumulated hierarchy level does not exceed D constitute the three-dimensional sparse B-spline basis \(\tilde{\varvec{{\Phi }}}^{(d,D)}(\mathbf {u}_1,\mathbf {u}_2,\mathbf {u}_3)\). Figure 3 shows the construction principle for a bivariate sparse B-spline basis with hierarchy level \(d=2\). Figure 4 presents the placements of the knots for the full tensor product of B-splines and the sparse B-spline basis for \(d=2\). The corresponding spline coefficients are given by \(\tilde{\mathbf {b}}^{(d,D)} = \mathcal{E(O_D)}^\top \tilde{\mathbf {b}}^{(d)}\).
Appendix 2: Marginal likelihood
The prior (13) is degenerated, which needs to be corrected as follows. For simplicity, we write \(\mathbf {b}:= \tilde{\mathbf {b}^{(d,D)}}\) in this section. We decompose \(\mathbf {b}\) into the two components \(\mathbf {b}_{\sim }\) and \(\mathbf {b}_{\bot }\), respectively, such that \(\mathbf {b}_{\sim }\) is a normally distributed random vector with non-degenerated variance and \(\mathbf {b}_{\bot }\) contains the remaining components treated as parameters, see also Wand and Ormerod (2008). Applying a singular value decomposition, we have
where \(\tilde{\mathbf {\Lambda }}\) is a diagonal matrix with positive eigenvalues and \(\tilde{\mathbf {U}} \in \mathbb {R}^{(K+1) \times h}\) are the corresponding eigenvectors with \(K+1\) being the number of elements in \(\mathbf {b}\) and \(h=K+1-4\) being the rank of \(\tilde{\mathbf {P}^{(D)}}\). Extending \(\tilde{\mathbf {U}}\) to an orthogonal basis by \(\check{\mathbf {U}}\) gives \(\mathbf {b}_{\sim }=\tilde{\mathbf {U}}^T\mathbf {b}\). With the a priori assumption \(\mathbf {b}_{\sim } \sim N(0, \lambda ^{-1}\tilde{\mathbf {\Lambda }}^{-1})\) and \(\mathbf {U}=(\tilde{\mathbf {U}}, \check{\mathbf {U}})\) as orthogonal basis, we get \(\mathbf {b}_{\bot }={\check{\mathbf {U}}}^T \mathbf {b}\). Conditioning on \(\mathbf {b}_{\sim }\), we get the mixed model log-likelihood
The integral can be approximated by a Laplace approximation (see also Rue et al. 2009)
where \(\hat{\mathbf {b}}\) denotes the penalized maximum likelihood estimate. We can now differentiate (25) with respect to \(\lambda \) which gives
and \({\mathbf H}_{pen}^{(d,D)}(\hat{\mathbf {b}},\lambda )\) denotes the second-order partial derivative of (10) with respect to \(\hat{\mathbf {b}}\), i.e.,
Appendix 3: Computing time
We present some computing times, defined as elapsed time on the system, measured by R on a machine with an Intel Core i7-2600 CPU @3.40 Ghz \(\times \) 4 using R.3.3.1 on Linux Mint 17.2 (64-bit). To this end, we take the first ten data sets of the three-dimensional non-simplified vine copulas and the five-dimensional non-simplified vine copulas from our simulation study, each for case (b) with \(\beta =0.6\) for both sample sizes \(n=500\) and \(n=2000\). Computing time is measured for (i) bivariate unconditional copula densities \(c_{12}\) for the first two marginal arguments from the selected data, (ii) bivariate conditional copula densities \(c_{12|3}\) with one conditioning argument and (iii) the estimation of the non-simplified vine copula. The computing times are realized with one starting value of \(\lambda \). Choosing three starting values for each copula estimation as in the simulation study increases the computing time linearly. See Table 3 for the results. The five-dimensional non-simplified vine copulas are computed in approximately 3–7 min, depending on the basis size and sample size, while the computing time for a three-dimensional non-simplified vine copula is less than 1 min. It might seem counterintuitive that the computing time decreases if the sample size increases from \(n=500\) to \(n=2000\) observations. That is because less iteration steps in the quadratic programming are required.
Rights and permissions
About this article
Cite this article
Schellhase, C., Spanhel, F. Estimating non-simplified vine copulas using penalized splines. Stat Comput 28, 387–409 (2018). https://doi.org/10.1007/s11222-017-9737-7
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11222-017-9737-7