Author:
Geenens Gery,de Micheaux Pierre Lafaye
Abstract
The Hellinger correlation is a dependence measure based on the Hellinger distance between the joint distribution of the bivariate random vector of interest and the distribution with the same marginals but showing independence. It is shown to enjoy many desirable properties, including two original ones shown to imply other usual postulates. A simple and efficient nonparametric estimator for that quantity is proposed, with its implementation publicly available in the R package HellCor.