|
Activity Number:
|
472
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Business and Economics Statistics Section
|
| Abstract - #310255 |
|
Title:
|
Bayesian Wavelet Estimation of Copulas for Dependence Modeling
|
|
Author(s):
|
Leming Qu*+
|
|
Companies:
|
Boise State University
|
|
Address:
|
Department of Mathematics, Boise, ID, 83725-1555,
|
|
Keywords:
|
Wavelet ; Bayesian ; Copulas ; MCMC ; Dependence Modeling
|
|
Abstract:
|
Copulas are full measures of dependence among random variables. It is now widely used in financial engineering for modeling high-dimensional problems, such as value-at-risk or portfolio credit risk. A copular's hidden dependence structure that couples a joint distribution with its marginals makes a parametric copular model non-trivial. We propose a nonparametric estimator using wavelet in the Bayesian framework. A mixture prior of a point mass at zero and a normal distribution is imposed on the wavelet coefficients. The Markov chain Monte Carlo (MCMC) algorithm is used for posterior inference. Performances are evaluated on simulated data and on a real dataset.
|
