IDEAS home Printed from https://ideas.repec.org/a/eee/csdana/v161y2021ics0167947321000876.html
   My bibliography  Save this article

Harmless label noise and informative soft-labels in supervised classification

Author

Listed:
  • Ahfock, Daniel
  • McLachlan, Geoffrey J.

Abstract

Manual labelling of training examples is common practice in supervised learning. When the labelling task is of non-trivial difficulty, the supplied labels may not be equal to the ground-truth labels, and label noise is introduced into the training dataset. If the manual annotation is carried out by multiple experts, the same training example can be given different class assignments by different experts, which is indicative of label noise. In the framework of model-based classification, a simple, but key observation is that when the manual labels are sampled using the posterior probabilities of class membership, the noisy labels are as valuable as the ground-truth labels in terms of statistical information. A relaxation of this process is a random effects model for imperfect labelling by a group that uses approximate posterior probabilities of class membership. The relative efficiency of logistic regression using the noisy labels compared to logistic regression using the ground-truth labels can then be derived. The main finding is that logistic regression can be robust to label noise when label noise and classification difficulty are positively correlated. In particular, when classification difficulty is the only source of label errors, multiple sets of noisy labels can supply more information for the estimation of a classification rule compared to the single set of ground-truth labels.

Suggested Citation

  • Ahfock, Daniel & McLachlan, Geoffrey J., 2021. "Harmless label noise and informative soft-labels in supervised classification," Computational Statistics & Data Analysis, Elsevier, vol. 161(C).
  • Handle: RePEc:eee:csdana:v:161:y:2021:i:c:s0167947321000876
    DOI: 10.1016/j.csda.2021.107253
    as

    Download full text from publisher

    File URL: http://www.sciencedirect.com/science/article/pii/S0167947321000876
    Download Restriction: Full text for ScienceDirect subscribers only.

    File URL: https://libkey.io/10.1016/j.csda.2021.107253?utm_source=ideas
    LibKey link: if access is restricted and if your library uses this service, LibKey will redirect you to where you can use your library subscription to access this item
    ---><---

    As the access to this document is restricted, you may want to search for a different version of it.

    References listed on IDEAS

    as
    1. A. P. Dawid & A. M. Skene, 1979. "Maximum Likelihood Estimation of Observer Error‐Rates Using the EM Algorithm," Journal of the Royal Statistical Society Series C, Royal Statistical Society, vol. 28(1), pages 20-28, March.
    2. Timothy I Cannings & Yingying Fan & Richard J Samworth, 2020. "Classification with imperfect training labels," Biometrika, Biometrika Trust, vol. 107(2), pages 311-330.
    3. Bi, Yingtao & Jeske, Daniel R., 2010. "The efficiency of logistic regression compared to normal discriminant analysis under class-conditional classification noise," Journal of Multivariate Analysis, Elsevier, vol. 101(7), pages 1622-1637, August.
    4. R. C. H. Cheng & W. B. Liu, 2001. "The Consistency of Estimators in Finite Mixture Models," Scandinavian Journal of Statistics, Danish Society for Theoretical Statistics;Finnish Statistical Society;Norwegian Statistical Association;Swedish Statistical Association, vol. 28(4), pages 603-616, December.
    5. Benjamin Quost & Thierry Denœux & Shoumei Li, 2017. "Parametric classification with soft labels using the evidential EM algorithm: linear discriminant analysis versus logistic regression," Advances in Data Analysis and Classification, Springer;German Classification Society - Gesellschaft für Klassifikation (GfKl);Japanese Classification Society (JCS);Classification and Data Analysis Group of the Italian Statistical Society (CLADAG);International Federation of Classification Societies (IFCS), vol. 11(4), pages 659-690, December.
    Full references (including those not matched with items on IDEAS)

    Most related items

    These are the items that most often cite the same works as this one and are cited by the same works as this one.
    1. Huiqing Chang & Linlin Huang & Panpan Song & Liyang Ru, 2022. "Prediction of arsenic accumulation in a calcareous soil-wheat/maize rotation system with continuous amendment of sewage sludge," Plant, Soil and Environment, Czech Academy of Agricultural Sciences, vol. 68(11), pages 516-524.
    2. Baiyun Chen & Longhai Huang & Zizhong Chen & Guoyin Wang, 2022. "An Ensemble and Iterative Recovery Strategy Based k GNN Method to Edit Data with Label Noise," Mathematics, MDPI, vol. 10(15), pages 1-28, August.
    3. Xiaoxiao Yang & Jing Zhang & Jun Peng & Lihong Lei, 2021. "Incentive mechanism based on Stackelberg game under reputation constraint for mobile crowdsensing," International Journal of Distributed Sensor Networks, , vol. 17(6), pages 15501477211, June.
    4. Junming Yin & Jerry Luo & Susan A. Brown, 2021. "Learning from Crowdsourced Multi-labeling: A Variational Bayesian Approach," Information Systems Research, INFORMS, vol. 32(3), pages 752-773, September.
    5. F. Bartolucci & A. Farcomeni & F. Pennoni, 2014. "Latent Markov models: a review of a general framework for the analysis of longitudinal data with covariates," TEST: An Official Journal of the Spanish Society of Statistics and Operations Research, Springer;Sociedad de Estadística e Investigación Operativa, vol. 23(3), pages 433-465, September.
    6. Yuqing Kong, 2021. "Information Elicitation Meets Clustering," Papers 2110.00952, arXiv.org.
    7. Tomer Geva & Maytal Saar‐Tsechansky, 2021. "Who Is a Better Decision Maker? Data‐Driven Expert Ranking Under Unobserved Quality," Production and Operations Management, Production and Operations Management Society, vol. 30(1), pages 127-144, January.
    8. Jesus Cerquides & Mehmet Oğuz Mülâyim & Jerónimo Hernández-González & Amudha Ravi Shankar & Jose Luis Fernandez-Marquez, 2021. "A Conceptual Probabilistic Framework for Annotation Aggregation of Citizen Science Data," Mathematics, MDPI, vol. 9(8), pages 1-15, April.
    9. Tin Lok James Ng & Thomas Brendan Murphy, 2021. "Model-based Clustering of Count Processes," Journal of Classification, Springer;The Classification Society, vol. 38(2), pages 188-211, July.
    10. Pennoni, Fulvia & Romeo, Isabella, 2016. "Latent Markov and growth mixture models for ordinal individual responses with covariates: a comparison," MPRA Paper 72939, University Library of Munich, Germany.
    11. Xiu Fang & Suxin Si & Guohao Sun & Quan Z. Sheng & Wenjun Wu & Kang Wang & Hang Lv, 2022. "Selecting Workers Wisely for Crowdsourcing When Copiers and Domain Experts Co-exist," Future Internet, MDPI, vol. 14(2), pages 1-22, January.
    12. Alaa Ghanaiem & Evgeny Kagan & Parteek Kumar & Tal Raviv & Peter Glynn & Irad Ben-Gal, 2023. "Unsupervised Classification under Uncertainty: The Distance-Based Algorithm," Mathematics, MDPI, vol. 11(23), pages 1-19, November.
    13. Yoshikazu Terada, 2014. "Strong Consistency of Reduced K-means Clustering," Scandinavian Journal of Statistics, Danish Society for Theoretical Statistics;Finnish Statistical Society;Norwegian Statistical Association;Swedish Statistical Association, vol. 41(4), pages 913-931, December.
    14. Yoichi Miyata & Takayuki Shiohama & Toshihiro Abe, 2020. "Estimation of finite mixture models of skew-symmetric circular distributions," Metrika: International Journal for Theoretical and Applied Statistics, Springer, vol. 83(8), pages 895-922, November.
    15. Jing Wang & Panagiotis G. Ipeirotis & Foster Provost, 2017. "Cost-Effective Quality Assurance in Crowd Labeling," Information Systems Research, INFORMS, vol. 28(1), pages 137-158, March.

    Corrections

    All material on this site has been provided by the respective publishers and authors. You can help correct errors and omissions. When requesting a correction, please mention this item's handle: RePEc:eee:csdana:v:161:y:2021:i:c:s0167947321000876. See general information about how to correct material in RePEc.

    If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about.

    If CitEc recognized a bibliographic reference but did not link an item in RePEc to it, you can help with this form .

    If you know of missing items citing this one, you can help us creating those links by adding the relevant references in the same way as above, for each refering item. If you are a registered author of this item, you may also want to check the "citations" tab in your RePEc Author Service profile, as there may be some citations waiting for confirmation.

    For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: Catherine Liu (email available below). General contact details of provider: http://www.elsevier.com/locate/csda .

    Please note that corrections may take a couple of weeks to filter through the various RePEc services.

    IDEAS is a RePEc service. RePEc uses bibliographic data supplied by the respective publishers.