Nonparametric kernel estimation of conditional copula density

Abstract

This paper introduces a new mathematical formula for the bivariate conditional copula density and proposes kernel-type estimators for it. We demonstrate the consistency and asymptotic normality of these estimators, which also exhibit the best quadratic mean convergence rate when the optimal theoretical bandwidth is selected.