Non and semiparametric methods based on copula functions

点击翻译按钮获取中文摘要

Title: Non and semiparametric methods based on copula functions We propose the development of new statistical methods based on copula functions for complete and right censored data. For censored data, we propose and investigate new, non and semiparametric, methods of estimation of the conditional dependence and of the regression function using copulas. Also, based on copula functions, we seek to design new tests of independence and conditional independence. A test of causality, which a very popular in economics and finance, is derived. Novel goodness-of-fit tests for unconditional and conditional copulas are explored.