Robust Bayesian Classification Using An Optimistic Score Ratio

Viet Anh Nguyen, Nian Si, Jose Blanchet
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:7327-7337, 2020.

Abstract

We build a Bayesian contextual classification model using an optimistic score ratio for robust binary classification when there is limited information on the class-conditional, or contextual, distribution. The optimistic score searches for the distribution that is most plausible to explain the observed outcomes in the testing sample among all distributions belonging to the contextual ambiguity set which is prescribed using a limited structural constraint on the mean vector and the covariance matrix of the underlying contextual distribution. We show that the Bayesian classifier using the optimistic score ratio is conceptually attractive, delivers solid statistical guarantees and is computationally tractable. We showcase the power of the proposed optimistic score ratio classifier on both synthetic and empirical data.

Cite this Paper


BibTeX
@InProceedings{pmlr-v119-nguyen20e, title = {Robust {B}ayesian Classification Using An Optimistic Score Ratio}, author = {Nguyen, Viet Anh and Si, Nian and Blanchet, Jose}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {7327--7337}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/nguyen20e/nguyen20e.pdf}, url = {https://proceedings.mlr.press/v119/nguyen20e.html}, abstract = {We build a Bayesian contextual classification model using an optimistic score ratio for robust binary classification when there is limited information on the class-conditional, or contextual, distribution. The optimistic score searches for the distribution that is most plausible to explain the observed outcomes in the testing sample among all distributions belonging to the contextual ambiguity set which is prescribed using a limited structural constraint on the mean vector and the covariance matrix of the underlying contextual distribution. We show that the Bayesian classifier using the optimistic score ratio is conceptually attractive, delivers solid statistical guarantees and is computationally tractable. We showcase the power of the proposed optimistic score ratio classifier on both synthetic and empirical data.} }
Endnote
%0 Conference Paper %T Robust Bayesian Classification Using An Optimistic Score Ratio %A Viet Anh Nguyen %A Nian Si %A Jose Blanchet %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-nguyen20e %I PMLR %P 7327--7337 %U https://proceedings.mlr.press/v119/nguyen20e.html %V 119 %X We build a Bayesian contextual classification model using an optimistic score ratio for robust binary classification when there is limited information on the class-conditional, or contextual, distribution. The optimistic score searches for the distribution that is most plausible to explain the observed outcomes in the testing sample among all distributions belonging to the contextual ambiguity set which is prescribed using a limited structural constraint on the mean vector and the covariance matrix of the underlying contextual distribution. We show that the Bayesian classifier using the optimistic score ratio is conceptually attractive, delivers solid statistical guarantees and is computationally tractable. We showcase the power of the proposed optimistic score ratio classifier on both synthetic and empirical data.
APA
Nguyen, V.A., Si, N. & Blanchet, J.. (2020). Robust Bayesian Classification Using An Optimistic Score Ratio. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:7327-7337 Available from https://proceedings.mlr.press/v119/nguyen20e.html.

Related Material