RUO Home

Repositorio Institucional de la Universidad de Oviedo

View Item 
  •   RUO Home
  • Producción Bibliográfica de UniOvi: RECOPILA
  • Artículos
  • View Item
  •   RUO Home
  • Producción Bibliográfica de UniOvi: RECOPILA
  • Artículos
  • View Item
    • español
    • English
JavaScript is disabled for your browser. Some features of this site may not work without it.

Browse

All of RUOCommunities and CollectionsBy Issue DateAuthorsTitlesSubjectsxmlui.ArtifactBrowser.Navigation.browse_issnAuthor profilesThis CollectionBy Issue DateAuthorsTitlesSubjectsxmlui.ArtifactBrowser.Navigation.browse_issn

My Account

LoginRegister

Statistics

View Usage Statistics

RECENTLY ADDED

Last submissions
Repository
How to publish
Resources
FAQs

An overview of Inference Methods in Probabilistic Classifier Chains for Multi-label classification

Author:
Mena Waldo, DeinerUniovi authority; Montañés Roces, ElenaUniovi authority; Quevedo Pérez, José RamónUniovi authority; Coz Velasco, Juan José delUniovi authority
Publication date:
2016
Editorial:

Wiley

Publisher version:
http://dx.doi.org/10.1002/widm.1185
Citación:
WIREs Data Mining and Knowledge Discovery, 6, p. 215-230 (2016); doi:10.1002/widm.1185
Descripción física:
p. 215-230
Abstract:

This study presents a review of the recent advances in performing inference in probabilistic classifier chains for multilabel classification. The interest of performing such inference arises in an attempt of improving the performance of the approach based on greedy search (the well‐known CC method) and simultaneously reducing the computational cost of an exhaustive search (the well‐known PCC method). Unlike PCC and as CC, inference techniques do not explore all the possible solutions, but they increase the performance of CC, sometimes reaching the optimal solution in terms of subset 0/1 loss, as PCC does. The ε‐approximate algorithm, the method based on a beam search and Monte Carlo sampling are those techniques. An exhaustive set of experiments over a wide range of datasets are performed to analyze not only to which extent these techniques tend to produce optimal solutions, otherwise also to study their computational cost, both in terms of solutions explored and execution time. Only ε‐approximate algorithm with ε=.0 theoretically guarantees reaching an optimal solution in terms of subset 0/1 loss. However, the other algorithms provide solutions close to an optimal solution, despite the fact they do not guarantee to reach an optimal solution. The ε‐approximate algorithm is the most promising to balance the performance in terms of subset 0/1 loss against the number of solutions explored and execution time. The value of ε determines a degree to which one prefers to guarantee to reach an optimal solution at the expense of increasing the computational cost

This study presents a review of the recent advances in performing inference in probabilistic classifier chains for multilabel classification. The interest of performing such inference arises in an attempt of improving the performance of the approach based on greedy search (the well‐known CC method) and simultaneously reducing the computational cost of an exhaustive search (the well‐known PCC method). Unlike PCC and as CC, inference techniques do not explore all the possible solutions, but they increase the performance of CC, sometimes reaching the optimal solution in terms of subset 0/1 loss, as PCC does. The ε‐approximate algorithm, the method based on a beam search and Monte Carlo sampling are those techniques. An exhaustive set of experiments over a wide range of datasets are performed to analyze not only to which extent these techniques tend to produce optimal solutions, otherwise also to study their computational cost, both in terms of solutions explored and execution time. Only ε‐approximate algorithm with ε=.0 theoretically guarantees reaching an optimal solution in terms of subset 0/1 loss. However, the other algorithms provide solutions close to an optimal solution, despite the fact they do not guarantee to reach an optimal solution. The ε‐approximate algorithm is the most promising to balance the performance in terms of subset 0/1 loss against the number of solutions explored and execution time. The value of ε determines a degree to which one prefers to guarantee to reach an optimal solution at the expense of increasing the computational cost

URI:
http://hdl.handle.net/10651/39325
ISSN:
1942-4787
DOI:
10.1002/widm.1185
Patrocinado por:

This research has been supported by the Spanish Ministerio de Economía y Competitividad (grants TIN2011-23558, TIN2015-65069)

Collections
  • Artículos [37541]
  • Informática [875]
  • Investigaciones y Documentos OpenAIRE [8401]
Files in this item
Thumbnail
untranslated
Postprint (821.4Kb)
Métricas
Compartir
Exportar a Mendeley
Estadísticas de uso
Estadísticas de uso
Metadata
Show full item record
Página principal Uniovi

Biblioteca

Contacto

Facebook Universidad de OviedoTwitter Universidad de Oviedo
The content of the Repository, unless otherwise specified, is protected with a Creative Commons license: Attribution-Non Commercial-No Derivatives 4.0 Internacional
Creative Commons Image