B1002
Title: Nonparametric estimation for extreme-value copula functions via constrained spline regressions
Authors: Yichen Qin - University of Cincinnati (United States) [presenting]
Yang Li - Renmin University of China (China)
Jun Yan - University of Connecticut (United States)
Siqi Xiang - Renmin University of China (China)
Abstract: A new nonparametric estimation procedure is introduced for extreme-value copulas using spline regressions. By fitting a shape constrained spline regression function to the data points obtained from the rank-based transformation of the original observations, the authors provide new estimates of the Pickands dependence functions of the extreme-value copula. In order to impose the shape constraints on the spline regression, a new set of basis functions which satisfies such constraints is proposed. Compared with existing methods, the method works well in simulation and in real data analysis.
Journal