2018/01/29 by Thomas Nagler, Nagler, Thomas, Christian Bumann +3 · 4 citations
Economics, Econometrics and Finance · Mathematics · Social Sciences · #Computation (stat.CO) #FOS: Computer and information sciences #Financial Risk and Volatility Modeling #Insurance, Mortality, Demography, Risk Management #Methodology (stat.ME) #Statistical Methods and Inference
paper · pdf · doi:10.48550/arxiv.1801.09739
openalex publication_date 2018/01/29 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Vine copulas allow to build flexible dependence models for an arbitrary\nnumber of variables using only bivariate building blocks. The number of\nparameters in a vine copula model increases quadratically with the dimension,\nwhich poses new challenges in high-dimensional applications. To alleviate the\ncomputational burden and risk of overfitting, we propose a modified Bayesian\ninformation criterion (BIC) tailored to sparse vine copula models. We show that\nthe criterion can consistently distinguish between the true and alternative\nmodels under less stringent conditions than the classical BIC. The new\ncriterion can be used to select the hyper-parameters of sparse model classes,\nsuch as truncated and thresholded vine copulas. We propose a computationally\nefficient implementation and illustrate the benefits of the new concepts with a\ncase study where we model the dependence in a large stock stock portfolio.\n