Correcting Sample Selection Bias by Unlabeled Data

Jiayuan Huang; Alexander J. Smola; Arthur Gretton; Karsten Borgwardt; Bernhard Schölkopf

doi:10.7551/mitpress/7503.003.0080

Abstract

We consider the scenario where training and test data are drawn from different distributions, commonly referred to as sample selection bias.Most algorithms for this setting try to first recover sampling distributions and then make appropriate corrections based on the distribution estimate.We present a nonparametric method which directly produces resampling weights without distribution estimation.Our method works by matching distributions between training and testing sets in feature space.Experimental results demonstrate that our method works well in practice.

Keywords

Selection biasSelection (genetic algorithm)Sample (material)Sampling biasComputer scienceArtificial intelligenceStatisticsPattern recognition (psychology)Sample size determinationMathematicsChromatographyChemistry

Affiliated Institutions

Related Publications

Arcing classifier (with discussion and a rejoinder by the author)

Leo Breiman

Recent work has shown that combining multiple versions of unstable\nclassifiers such as trees or neural nets results in reduced test set error. One\nof the more effective is bag...

1998 The Annals of Statistics 1088 citations

Moment Matching for Multi-Source Domain Adaptation

Xingchao Peng , Qinxun Bai , Xide Xia +3 more

Conventional unsupervised domain adaptation (UDA) assumes that training data are sampled from a single domain. This neglects the more practical scenario where training data are ...

2019 2019 IEEE/CVF International Conferenc... 1443 citations

Deep Convolutional Transfer Learning Network: A New Method for Intelligent Fault Diagnosis of Machines With Unlabeled Data

Liang Guo , Yaguo Lei , Saibo Xing +2 more

The success of intelligent fault diagnosis of machines relies on the following two conditions: 1) labeled data with fault information are available; and 2) the training and test...

2018 IEEE Transactions on Industrial Elect... 1138 citations

Boosting the margin: a new explanation for the effectiveness of voting methods

Peter L. Bartlett , Yoav Freund , Wee Sun Lee +1 more

One of the surprising recurring phenomena observed in experiments\nwith boosting is that the test error of the generated classifier usually does\nnot increase as its size become...

1998 The Annals of Statistics 2310 citations

Exact Convex Confidence-Weighted Learning

Koby Crammer , Mark Dredze , Fernando Pereira

Confidence-weighted (CW) learning [6], an online learning method for linear clas-sifiers, maintains a Gaussian distributions over weight vectors, with a covariance matrix that r...

2008 126 citations

Publication Info

Year: 2007
Type: book-chapter
Pages: 601-608
Citations: 1529
Access: Closed

External Links

Download PDF (Free) View on DOI.org Semantic Scholar

Social Impact

Altmetric

Correcting Sample Selection Bias by Unlabeled Data

PlumX Metrics

Social media, news, blog, policy document mentions

Citation Metrics

1529

OpenAlex

213

Influential

415

CrossRef

Cite This

APA Style

                            
                                    Jiayuan Huang, 
                                
                                    Alexander J. Smola, 
                                
                                    Arthur Gretton
                                
                                et al.
                            
                            (2007). 
                            Correcting Sample Selection Bias by Unlabeled Data. 
                            The MIT Press eBooks
                            
                            , 601-608.
                            https://doi.org/10.7551/mitpress/7503.003.0080

Identifiers

DOI: 10.7551/mitpress/7503.003.0080

Data Quality

Data completeness: 81%