ABSTRACT

For estimation and simulation from high-dimensional copulas, such as vines and factor models, and for converting from one parametrization to another for Gaussian dependence models, there are sequential procedures that are best summarized as algorithms. In this chapter, the algorithms are presented as pseudo-code1. Algorithms include mappings from partial correlations to correlations and/or regression coefficients for Gaussian truncated vines and factor models. This is important because Section 5.1 suggests to initially fit and compare different dependence structures based on an assumption of multivariate Gaussian copulas.