Mixture density network estimation of continuous variable maximum likelihood using discrete training samples

Abstract Mixture density networks (MDNs) can be used to generate posterior density functions of model parameters $$\varvec{\theta }$$ θ given a set of observables $${\mathbf {x}}$$ x . In some applications, training data are available only for discrete values of a continuous parameter $$\varvec{\the...

Full description

Bibliographic Details
Main Authors:	Charles Burton, Spencer Stubbs, Peter Onyisi
Format:	Article
Language:	English
Published:	SpringerOpen 2021-07-01
Series:	European Physical Journal C: Particles and Fields
Online Access:	https://doi.org/10.1140/epjc/s10052-021-09469-y

id	doaj-42b642e65c024959bb56db60384144f4
record_format	Article
spelling	doaj-42b642e65c024959bb56db60384144f42021-08-01T11:12:41ZengSpringerOpenEuropean Physical Journal C: Particles and Fields1434-60441434-60522021-07-018171910.1140/epjc/s10052-021-09469-yMixture density network estimation of continuous variable maximum likelihood using discrete training samplesCharles Burton0Spencer Stubbs1Peter Onyisi2Department of Physics, University of TexasDepartment of Physics, University of TexasDepartment of Physics, University of TexasAbstract Mixture density networks (MDNs) can be used to generate posterior density functions of model parameters $$\varvec{\theta }$$ θ given a set of observables $${\mathbf {x}}$$ x . In some applications, training data are available only for discrete values of a continuous parameter $$\varvec{\theta }$$ θ . In such situations, a number of performance-limiting issues arise which can result in biased estimates. We demonstrate the usage of MDNs for parameter estimation, discuss the origins of the biases, and propose a corrective method for each issue.https://doi.org/10.1140/epjc/s10052-021-09469-y
collection	DOAJ
language	English
format	Article
sources	DOAJ
author	Charles Burton Spencer Stubbs Peter Onyisi
spellingShingle	Charles Burton Spencer Stubbs Peter Onyisi Mixture density network estimation of continuous variable maximum likelihood using discrete training samples European Physical Journal C: Particles and Fields
author_facet	Charles Burton Spencer Stubbs Peter Onyisi
author_sort	Charles Burton
title	Mixture density network estimation of continuous variable maximum likelihood using discrete training samples
title_short	Mixture density network estimation of continuous variable maximum likelihood using discrete training samples
title_full	Mixture density network estimation of continuous variable maximum likelihood using discrete training samples
title_fullStr	Mixture density network estimation of continuous variable maximum likelihood using discrete training samples
title_full_unstemmed	Mixture density network estimation of continuous variable maximum likelihood using discrete training samples
title_sort	mixture density network estimation of continuous variable maximum likelihood using discrete training samples
publisher	SpringerOpen
series	European Physical Journal C: Particles and Fields
issn	1434-6044 1434-6052
publishDate	2021-07-01
description	Abstract Mixture density networks (MDNs) can be used to generate posterior density functions of model parameters $$\varvec{\theta }$$ θ given a set of observables $${\mathbf {x}}$$ x . In some applications, training data are available only for discrete values of a continuous parameter $$\varvec{\theta }$$ θ . In such situations, a number of performance-limiting issues arise which can result in biased estimates. We demonstrate the usage of MDNs for parameter estimation, discuss the origins of the biases, and propose a corrective method for each issue.
url	https://doi.org/10.1140/epjc/s10052-021-09469-y
work_keys_str_mv	AT charlesburton mixturedensitynetworkestimationofcontinuousvariablemaximumlikelihoodusingdiscretetrainingsamples AT spencerstubbs mixturedensitynetworkestimationofcontinuousvariablemaximumlikelihoodusingdiscretetrainingsamples AT peteronyisi mixturedensitynetworkestimationofcontinuousvariablemaximumlikelihoodusingdiscretetrainingsamples
_version_	1721246187107909632

Mixture density network estimation of continuous variable maximum likelihood using discrete training samples

Similar Items