Dynamic Bayesian combination of multiple imperfect classifiers

DOI10.1007/978-3-642-36406-8_1zbMATH Open1346.68164arXiv1206.1831OpenAlexW2157010107MaRDI QIDQ2822296FDOQ2822296

Authors: Edwin Simpson, Stephen J. Roberts 

Publication date: 30 September 2016

Published in: Decision Making and Imperfection (Search for Journal in Brave)

Abstract: Classifier combination methods need to make best use of the outputs of multiple, imperfect classifiers to enable higher accuracy classifications. In many situations, such as when human decisions need to be combined, the base decisions can vary enormously in reliability. A Bayesian approach to such uncertain combination allows us to infer the differences in performance between individuals and to incorporate any available prior knowledge about their abilities when training data is sparse. In this paper we explore Bayesian classifier combination, using the computationally efficient framework of variational Bayesian inference. We apply the approach to real data from a large citizen science project, Galaxy Zoo Supernovae, and show that our method far outperforms other established approaches to imperfect decision combination. We go on to analyse the putative community structure of the decision makers, based on their inferred decision making strategies, and show that natural groupings are formed. Finally we present a dynamic Bayesian classifier combination approach and investigate the changes in base classifier performance over time.

Full work available at URL: https://arxiv.org/abs/1206.1831

Recommendations

• A Bayesian framework for the combination of classifier outputs
• The combination of multiple classifiers using an evidential reasoning approach
• Robust naive Bayes combination of multiple classifications
• scientific article; zbMATH DE number 1860779
• Combining Classifiers via Discretization

Mathematics Subject Classification ID

Bayesian inference (62F15) Learning and adaptive systems in artificial intelligence (68T05) Pattern recognition, speech recognition (68T10)

Cites Work

• Title not available (Why is that?)
• Dynamic Generalized Linear Models and Bayesian Forecasting
• Pattern recognition and machine learning.
• On Information and Sufficiency
• Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images
• Community structure in social and biological networks
• Adaptive Rejection Sampling for Gibbs Sampling
• Bayesian forecasting and dynamic models.
• The weighted majority algorithm
• Statistical field theory. With a foreword by David Pines
• Bayesian Inference on Multiscale Models for Poisson Intensity Estimation: Applications to Photon-Limited Image Denoising
• Variational Mixture of Bayesian Independent Component Analyzers
• Title not available (Why is that?)
• Fixed interval smoothing with discrete measurements

Cited In (7)

• Bayesian Methods for Intelligent Task Assignment in Crowdsourcing Systems
• Quality-aware online task assignment mechanisms using latent topic model
• Learning from crowds with sparse and imbalanced annotations
• Title not available (Why is that?)
• A Bayesian framework for the combination of classifier outputs
• A joint Gaussian process model for active visual recognition with expertise estimation in crowdsourcing
• Approximate Bayesian recursive estimation

This page was built for publication: Dynamic Bayesian combination of multiple imperfect classifiers

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2822296)