|
Activity Number:
|
74
- Statistical Methods and Applications: Domestic and International
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, July 28, 2019 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Business and Economic Statistics Section
|
|
Abstract #304992
|
|
|
Title:
|
Nonparametric Estimation and Testing for Positively Quadrant Dependent Copula
|
|
Author(s):
|
Lu Lu* and Sujit Ghosh
|
|
Companies:
|
North Carolina State University and North Carolina State Univ.
|
|
Keywords:
|
copula;
nonparametric methods;
PQD;
Bernstein copula;
Anderson-Darling distance
|
|
Abstract:
|
The estimation and testing of positive quadrant dependent (PQD) copula has found many applications in the field of reliability and finance. Many parametric families have been used to model PQD but the literature still lacks a fully automated method for the estimation and testing of PQD. This paper provides a class of flexible models based on Bernstein polynomials to approximate any arbitrary continuous copula and proposes an Anderson-Darling based criteria for estimation of PQD copula. In addition, an algorithm to dynamically choose the degrees of the Bernstein copula is developed and a test for the null hypothesis of PQD is constructed.
|
Authors who are presenting talks have a * after their name.
