Skip to content

Research seminar about a fast selection technique for bivariate copulae.

Notifications You must be signed in to change notification settings

nikpau/cop-select

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Speed check: copula selection and parameter estimation using non-parametric moment inverses and maximum likelihood

The purpose of this article is to conduct a Monte-Carlo simulation study that compares the efficiency, accuracy and empirical computation time of three different estimators in terms of selecting copulae and estimating their parameters. I will compare two moment-based estimators, the inverse of Kendall’s Tau and Blomqvist’s Beta, to the classic canonical maximum likelihood estimator. I find that, in terms of estimating parameters for a given copula, the moment-based estimators are inferior to maximum likelihood, especially for small sample sizes. For selecting copulae from a set of data, however, all three estimators are equally accurate while the moment-based ones are at least 16 times faster computationally.

Main article

The main article is fast_copula_estimation.pdf

About

Research seminar about a fast selection technique for bivariate copulae.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published