1: \begin{abstract}

2: Distributional functionals are integral functionals of one or more

3: probability distributions. Distributional functionals include information

4: measures such as entropy, mutual information, and divergence. Recent

5: work has focused on the problem of nonparametric estimation of entropy

6: and information divergence functionals. Many existing approaches are

7: restrictive in their assumptions on the density support set or require

8: difficult calculations at the support boundary which must be known

9: \emph{a priori}. The MSE convergence rate of a leave-one-out kernel

10: density plug-in divergence functional estimator for general bounded

11: density support sets is derived where knowledge of the support boundary

12: is not required. The theory of optimally weighted ensemble estimation

13: is generalized to derive two estimators that achieve the parametric

14: rate when the densities are sufficiently smooth. The asymptotic distribution

15: of these estimators and some guidelines for tuning parameter selection

16: are provided. Based on the theory, an empirical estimator of R\'enyi-$\alpha$

17: divergence is proposed that outperforms the standard kernel density

18: plug-in estimator, especially in high dimension. The estimators are

19: shown to be robust to the choice of tuning parameters.

20: \end{abstract}

21: