Bayesian Estimation of Generalized Partition of Unity Copulas
Andreas Masuhr · RePEc: Research Papers in Economics · 2018
This paper proposes two (Metropolis-Hastings) algorithms to estimate Generalized Partition of Unity Copulas (GPUC), a new class of nonparametric copulas that includes the versatile Bernstein Copula as a special case. Additionally a prior distribution for the parameter Matrix of GPUCs is established via Importance Sampling and an algorithm to sample such matrices is introduced. Finally, simulation studies show the effectiveness of the presented algorithms.