Consistent algorithms for multiclass classification with an abstain option

DOI10.1214/17-EJS1388MaRDI QIDQ1697489zbMATH OpenOpenAlexWikidataFDO

Authors Harish G. Ramaswamy, Ambuj Tewari, Shivani Agarwal

Publication date 20 February 2018

Published in Electronic Journal of Statistics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1505.04137, https://projecteuclid.org/euclid.ejs/1519030879

zbMATH Keywords

classificationmachine learningcalibrationsurrogate lossstatistical consistencyabstain loss

Mathematics Subject Classification ID

Classification and discrimination; cluster analysis (statistical aspects) (62H30) Pattern recognition, speech recognition (68T10)

Abstract: We consider the problem of $n$-class classification ($ngeq2$), where the classifier can choose to abstain from making predictions at a given cost, say, a factor $alpha$ of the cost of misclassification. Designing consistent algorithms for such $n$-class classification problems with a `reject option' is the main goal of this paper, thereby extending and generalizing previously known results for $n=2$. We show that the Crammer-Singer surrogate and the one vs all hinge loss, albeit with a different predictor than the standard argmax, yield consistent algorithms for this problem when $alpha=frac12$. More interestingly, we design a new convex surrogate that is also consistent for this problem when $alpha=frac12$ and operates on a much lower dimensional space ($log(n)$ as opposed to $n$). We also generalize all three surrogates to be consistent for any $alphain[0,frac12]$.

Recommendations

• Classification methods with reject option based on convex risk minimization
• Classification with a reject option using a hinge loss
• Classification with reject option
• On reject and refine options in multicategory classification
• Multilabel classification with partial abstention: Bayes-optimal prediction under label independence

Cites work

• scientific article; zbMATH DE number 6378119 (Why is no real title available?)
• scientific article; zbMATH DE number 1759592 (Why is no real title available?)
• scientific article; zbMATH DE number 2089926 (Why is no real title available?)
• scientific article; zbMATH DE number 823069 (Why is no real title available?)
• 10.1162/15324430260185628
• Analysis of error-reject trade-off in linearly combined multiple classifiers
• Classification methods with reject option based on convex risk minimization
• Classification using geometric level sets
• Classification with a reject option using a hinge loss
• Composite multiclass losses
• In defense of one-vs-all classification
• Learning with Rejection
• Multicategory Support Vector Machines
• On optimum recognition error and reject tradeoff
• On reject and refine options in multicategory classification
• On the consistency of multiclass classification methods
• On the foundations of noise-free selective classification
• Reducing multiclass to binary: A unifying approach for margin classifiers
• Statistical analysis of some multi-category large margin classification methods
• Structural, Syntactic, and Statistical Pattern Recognition

Cited in

(7)

• scientific article; zbMATH DE number 7306865 (Why is no real title available?)
• Efficient set-valued prediction in multi-class classification
• Aleatoric and epistemic uncertainty in machine learning: an introduction to concepts and methods
• Multilabel classification with partial abstention: Bayes-optimal prediction under label independence
• Multi-stage classifier design
• Least Ambiguous Set-Valued Classifiers With Bounded Error Levels
• Optimal rates for nonparametric F-score binary classification via post-processing

Describes a project that uses

Uses Software

• SVMlight

This page was built for publication: Consistent algorithms for multiclass classification with an abstain option

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1697489)