Efficient maximum likelihood estimation of copula based meta t-distributions
Recently an efficient fixed point algorithm, called maximization by parts (MBP), for finding maximum likelihood estimates has been applied to models based on Gaussian copulas. It requires a decomposition of a likelihood function into two parts and their iterative maximization by solving score equations. For the first time, the MBP algorithm is applied to multivariate meta t-distributions based on t-copulas. Since score equations for meta t-distributions do not have closed forms the proposed MBP algorithm in two variations maximizes the decomposed parts of the likelihood iteratively. Superiority of the proposed MBP algorithm over standard estimation methods such as inference for margins and direct maximization is illustrated in a simulation study. The usefulness of the proposed algorithm is shown in two data applications.
Year of publication: |
2011
|
---|---|
Authors: | Zhang, Ran ; Czado, Claudia ; Min, Aleksey |
Published in: |
Computational Statistics & Data Analysis. - Elsevier, ISSN 0167-9473. - Vol. 55.2011, 3, p. 1196-1214
|
Publisher: |
Elsevier |
Keywords: | Copula Inference for margins Maximum likelihood estimation Maximization by parts Meta-t distribution Rolling windows |
Saved in:
Online Resource
Saved in favorites
Similar items by person
-
Czado, Claudia, (2006)
-
Validating linear restrictions in linear regression models with general error structure
Holzmann, Hajo, (2006)
-
Testing for zero-modification in count regression models
Czado, Claudia, (2006)
- More ...