Efficient Monte Carlo methods for multi-dimensional learning with classifier chains

DOI10.1016/J.PATCOG.2013.10.006zbMATH Open1326.68251arXiv1211.2190OpenAlexW2006462052WikidataQ57253450 ScholiaQ57253450MaRDI QIDQ898396FDOQ898396

David Luengo, Luca Martino, Jesse Read

Publication date: 8 December 2015

Published in: Pattern Recognition (Search for Journal in Brave)

Abstract: Multi-dimensional classification (MDC) is the supervised learning problem where an instance is associated with multiple classes, rather than with a single class, as in traditional classification problems. Since these classes are often strongly correlated, modeling the dependencies between them allows MDC methods to improve their performance - at the expense of an increased computational cost. In this paper we focus on the classifier chains (CC) approach for modeling dependencies, one of the most popular and highest- performing methods for multi-label classification (MLC), a particular case of MDC which involves only binary classes (i.e., labels). The original CC algorithm makes a greedy approximation, and is fast but tends to propagate errors along the chain. Here we present novel Monte Carlo schemes, both for finding a good chain sequence and performing efficient inference. Our algorithms remain tractable for high-dimensional data sets and obtain the best predictive performance across several real data sets.

Full work available at URL: https://arxiv.org/abs/1211.2190

Recommendations

• Multi-dimensional classification with Bayesian networks
• An analysis of chaining in multi-label classification
• Scalable multi-output label prediction: from classifier chains to classifier trellises
• Inference and Learning in Multi-dimensional Bayesian Network Classifiers
• Bayesian network based label correlation analysis for multi-label classifier chain

zbMATH Keywords

Bayesian inferenceMonte Carlo methodsmulti-label classificationmulti-dimensional classificationclassifier chains

Mathematics Subject Classification ID

Classification and discrimination; cluster analysis (statistical aspects) (62H30) Monte Carlo methods (65C05) Pattern recognition, speech recognition (68T10)

Cites Work

• Optimization by Simulated Annealing
• Title not available (Why is that?)
• Multi-dimensional classification with Bayesian networks
• Foundations of computational intelligence. Volume 5: Function approximation and classification
• Classification by pairwise coupling
• On label dependence and loss minimization in multi-label classification

Cited In (9)

• Progressive random \(k\)-labelsets for cost-sensitive multi-label classification
• Using tensor products to detect unconditional label dependence in multilabel classifications
• Probability-based label enhancement for multi-dimensional classification
• A family of admissible heuristics for \(\mathrm{A}^*\) to perform inference in probabilistic classifier chains
• Title not available (Why is that?)
• Scalable multi-output label prediction: from classifier chains to classifier trellises
• Multi-target regression via input space expansion: treating targets as inputs
• Beam search algorithms for multilabel learning
• Classifier Chains: A Review and Perspectives

Uses Software

• WEKA
• MULAN

This page was built for publication: Efficient Monte Carlo methods for multi-dimensional learning with classifier chains

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q898396)