Publications

This paper presents a simple and intuitive index that attempts to summarize two non-parametric kernel density estimators in order to quantify the degree of convergence between both distributions. Its derivation follows a similar argument as the Gini index, in the sense that it can be expressed as a ratio of areas under the corresponding densities. The index can be routinely computed once kernel density estimators have been obtained, and it is bounded between 0, absence of convergence, and 1, full convergence. In this way it complements numerically the common graphical device of inspecting non-parametric kernel density estimators in exploratory data analysis.