Cargando…

AKL-ABC: An Automatic Approximate Bayesian Computation Approach Based on Kernel Learning

Bayesian statistical inference under unknown or hard to asses likelihood functions is a very challenging task. Currently, approximate Bayesian computation (ABC) techniques have emerged as a widely used set of likelihood-free methods. A vast number of ABC-based approaches have appeared in the literat...

Descripción completa

Detalles Bibliográficos
Autores principales: González-Vanegas, Wilson, Álvarez-Meza, Andrés, Hernández-Muriel, José, Orozco-Gutiérrez, Álvaro
Formato: Online Artículo Texto
Lenguaje:English
Publicado: MDPI 2019
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7514265/
http://dx.doi.org/10.3390/e21100932
_version_ 1783586548307984384
author González-Vanegas, Wilson
Álvarez-Meza, Andrés
Hernández-Muriel, José
Orozco-Gutiérrez, Álvaro
author_facet González-Vanegas, Wilson
Álvarez-Meza, Andrés
Hernández-Muriel, José
Orozco-Gutiérrez, Álvaro
author_sort González-Vanegas, Wilson
collection PubMed
description Bayesian statistical inference under unknown or hard to asses likelihood functions is a very challenging task. Currently, approximate Bayesian computation (ABC) techniques have emerged as a widely used set of likelihood-free methods. A vast number of ABC-based approaches have appeared in the literature; however, they all share a hard dependence on free parameters selection, demanding expensive tuning procedures. In this paper, we introduce an automatic kernel learning-based ABC approach, termed AKL-ABC, to automatically compute posterior estimations from a weighting-based inference. To reach this goal, we propose a kernel learning stage to code similarities between simulation and parameter spaces using a centered kernel alignment (CKA) that is automated via an Information theoretic learning approach. Besides, a local neighborhood selection (LNS) algorithm is used to highlight local dependencies over simulations relying on graph theory. Attained results on synthetic and real-world datasets show our approach is a quite competitive method compared to other non-automatic state-of-the-art ABC techniques.
format Online
Article
Text
id pubmed-7514265
institution National Center for Biotechnology Information
language English
publishDate 2019
publisher MDPI
record_format MEDLINE/PubMed
spelling pubmed-75142652020-11-09 AKL-ABC: An Automatic Approximate Bayesian Computation Approach Based on Kernel Learning González-Vanegas, Wilson Álvarez-Meza, Andrés Hernández-Muriel, José Orozco-Gutiérrez, Álvaro Entropy (Basel) Article Bayesian statistical inference under unknown or hard to asses likelihood functions is a very challenging task. Currently, approximate Bayesian computation (ABC) techniques have emerged as a widely used set of likelihood-free methods. A vast number of ABC-based approaches have appeared in the literature; however, they all share a hard dependence on free parameters selection, demanding expensive tuning procedures. In this paper, we introduce an automatic kernel learning-based ABC approach, termed AKL-ABC, to automatically compute posterior estimations from a weighting-based inference. To reach this goal, we propose a kernel learning stage to code similarities between simulation and parameter spaces using a centered kernel alignment (CKA) that is automated via an Information theoretic learning approach. Besides, a local neighborhood selection (LNS) algorithm is used to highlight local dependencies over simulations relying on graph theory. Attained results on synthetic and real-world datasets show our approach is a quite competitive method compared to other non-automatic state-of-the-art ABC techniques. MDPI 2019-09-24 /pmc/articles/PMC7514265/ http://dx.doi.org/10.3390/e21100932 Text en © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
spellingShingle Article
González-Vanegas, Wilson
Álvarez-Meza, Andrés
Hernández-Muriel, José
Orozco-Gutiérrez, Álvaro
AKL-ABC: An Automatic Approximate Bayesian Computation Approach Based on Kernel Learning
title AKL-ABC: An Automatic Approximate Bayesian Computation Approach Based on Kernel Learning
title_full AKL-ABC: An Automatic Approximate Bayesian Computation Approach Based on Kernel Learning
title_fullStr AKL-ABC: An Automatic Approximate Bayesian Computation Approach Based on Kernel Learning
title_full_unstemmed AKL-ABC: An Automatic Approximate Bayesian Computation Approach Based on Kernel Learning
title_short AKL-ABC: An Automatic Approximate Bayesian Computation Approach Based on Kernel Learning
title_sort akl-abc: an automatic approximate bayesian computation approach based on kernel learning
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7514265/
http://dx.doi.org/10.3390/e21100932
work_keys_str_mv AT gonzalezvanegaswilson aklabcanautomaticapproximatebayesiancomputationapproachbasedonkernellearning
AT alvarezmezaandres aklabcanautomaticapproximatebayesiancomputationapproachbasedonkernellearning
AT hernandezmurieljose aklabcanautomaticapproximatebayesiancomputationapproachbasedonkernellearning
AT orozcogutierrezalvaro aklabcanautomaticapproximatebayesiancomputationapproachbasedonkernellearning