The contribution of this paper is two-fold. First of all, we introduced an algorithm that estimates the joint density function using the smoothed recursive non-parametric estimator. Secondly, for the first time, we present a smoothed recursive non-parametric estimator for a multivariate copula density, a joint distribution that can capture complex dependence structures among its variates. Also, we present the numerical results that show an apparent convergence of these estimators.
This repository contains all necessary materials to replicate the results that were presented in the thesis.