Density estimation by dual ascent of the log-likelihood

A methodology is developed to assign, from an observed sample, a joint-probability distribution to a set of continuous variables. The algorithm proposed performs this assignment by mapping the original variables onto a jointly-Gaussian set. The map is built iteratively, ascending the log-likelihood of the observations, through a series of steps that move the marginal distributions along a random set of orthogonal directions towards normality.

Keywords

Density estimation, machine learning, maximum likelihood

2010 Mathematics Subject Classification

60H35, 65C30, 65L20

Full Text (PDF format)

Published 1 January 2010