Positive-unlabeled classification under class-prior shift: a prior-invariant approach based on density ratio estimation

DOI10.1007/S10994-022-06190-ZMaRDI QIDQ6106437zbMATH OpenOpenAlexFDO

Authors Shota Nakajima, Masashi Sugiyama

Publication date 27 June 2023

Published in Machine Learning (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2107.05045

zbMATH Keywords

class-prior shiftdensity ratio estimationpositive-unlabeled classification

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05)

Abstract: Learning from positive and unlabeled (PU) data is an important problem in various applications. Most of the recent approaches for PU classification assume that the class-prior (the ratio of positive samples) in the training unlabeled dataset is identical to that of the test data, which does not hold in many practical cases. In addition, we usually do not know the class-priors of the training and test data, thus we have no clue on how to train a classifier without them. To address these problems, we propose a novel PU classification method based on density ratio estimation. A notable advantage of our proposed method is that it does not require the class-priors in the training phase; class-prior shift is incorporated only in the test phase. We theoretically justify our proposed method and experimentally demonstrate its effectiveness.

Recommendations

• Principled analytic classifier for positive-unlabeled learning via weighted integral probability metric
• Information-theoretic representation learning for positive-unlabeled classification
• Class-prior estimation for learning from positive and unlabeled data
• A boosting framework for positive-unlabeled learning
• Learning from positive and unlabeled data: a survey

Cites work

• A least-squares approach to direct importance estimation
• Bipartite ranking: a risk-theoretic perspective
• Calibrated asymmetric surrogate losses
• Classification with asymmetric label noise: consistency and maximal denoising
• Convexity, Classification, and Risk Bounds
• Density ratio estimation in machine learning. Foreword by Thomas G. Dietterich
• Density-ratio matching under the Bregman divergence: a unified framework of density-ratio estimation
• Direct importance estimation for covariate shift adaptation
• How to compare different loss functions and their risks
• scientific article; zbMATH DE number 2080441 (Why is no real title available?)
• scientific article; zbMATH DE number 823069 (Why is no real title available?)
• Paradoxes of fuzzy logic, revisited
• Principal component analysis: a review and recent developments
• Ranking and empirical minimization of \(U\)-statistics
• Semi-supervised novelty detection
• Size-independent sample complexity of neural networks
• The hardness of approximate optima in lattices, codes, and systems of linear equations

Cited in

(3)

• Bayesian logistic model for positive and unlabeled data
• GKF-PUAL: a group kernel-free approach to positive-unlabeled learning with variable selection
• A boosting framework for positive-unlabeled learning

This page was built for publication: Positive-unlabeled classification under class-prior shift: a prior-invariant approach based on density ratio estimation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6106437)