作者: Irène Gijbels , Noël Veraverbeke , Marel Omelka
DOI: 10.1016/J.CSDA.2010.11.010
关键词:
摘要: One way to model a dependence structure is through the copula function which mean capture in joint distribution of variables. Association measures such as Kendall's tau or Spearman's rho can be expressed functionals copula. The between two variables highly influenced by covariate, and it real interest know how this changes with value taken covariate. This motivates need for introducing conditional copulas, associated association measures. After introduction motivation these concepts, nonparametric estimators are proposed discussed. Then estimates derived. A key issue that now looked at functions performances all investigated via simulation study also includes data-driven algorithm choosing smoothing parameters. usefulness methods illustrated on data examples.