Conditional Bernoulli Mixtures for Multi-label Classification

Cheng Li, Bingyu Wang, Virgil Pavlu, Javed Aslam
; Proceedings of The 33rd International Conference on Machine Learning, PMLR 48:2482-2491, 2016.

Abstract

Multi-label classification is an important machine learning task wherein one assigns a subset of candidate labels to an object. In this paper, we propose a new multi-label classification method based on Conditional Bernoulli Mixtures. Our proposed method has several attractive properties: it captures label dependencies; it reduces the multi-label problem to several standard binary and multi-class problems; it subsumes the classic independent binary prediction and power-set subset prediction methods as special cases; and it exhibits accuracy and/or computational complexity advantages over existing approaches. We demonstrate two implementations of our method using logistic regressions and gradient boosted trees, together with a simple training procedure based on Expectation Maximization. We further derive an efficient prediction procedure based on dynamic programming, thus avoiding the cost of examining an exponential number of potential label subsets. Experimental results show the effectiveness of the proposed method against competitive alternatives on benchmark datasets.

Cite this Paper


BibTeX
@InProceedings{pmlr-v48-lij16, title = {Conditional Bernoulli Mixtures for Multi-label Classification}, author = {Cheng Li and Bingyu Wang and Virgil Pavlu and Javed Aslam}, booktitle = {Proceedings of The 33rd International Conference on Machine Learning}, pages = {2482--2491}, year = {2016}, editor = {Maria Florina Balcan and Kilian Q. Weinberger}, volume = {48}, series = {Proceedings of Machine Learning Research}, address = {New York, New York, USA}, month = {20--22 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v48/lij16.pdf}, url = {http://proceedings.mlr.press/v48/lij16.html}, abstract = {Multi-label classification is an important machine learning task wherein one assigns a subset of candidate labels to an object. In this paper, we propose a new multi-label classification method based on Conditional Bernoulli Mixtures. Our proposed method has several attractive properties: it captures label dependencies; it reduces the multi-label problem to several standard binary and multi-class problems; it subsumes the classic independent binary prediction and power-set subset prediction methods as special cases; and it exhibits accuracy and/or computational complexity advantages over existing approaches. We demonstrate two implementations of our method using logistic regressions and gradient boosted trees, together with a simple training procedure based on Expectation Maximization. We further derive an efficient prediction procedure based on dynamic programming, thus avoiding the cost of examining an exponential number of potential label subsets. Experimental results show the effectiveness of the proposed method against competitive alternatives on benchmark datasets.} }
Endnote
%0 Conference Paper %T Conditional Bernoulli Mixtures for Multi-label Classification %A Cheng Li %A Bingyu Wang %A Virgil Pavlu %A Javed Aslam %B Proceedings of The 33rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2016 %E Maria Florina Balcan %E Kilian Q. Weinberger %F pmlr-v48-lij16 %I PMLR %J Proceedings of Machine Learning Research %P 2482--2491 %U http://proceedings.mlr.press %V 48 %W PMLR %X Multi-label classification is an important machine learning task wherein one assigns a subset of candidate labels to an object. In this paper, we propose a new multi-label classification method based on Conditional Bernoulli Mixtures. Our proposed method has several attractive properties: it captures label dependencies; it reduces the multi-label problem to several standard binary and multi-class problems; it subsumes the classic independent binary prediction and power-set subset prediction methods as special cases; and it exhibits accuracy and/or computational complexity advantages over existing approaches. We demonstrate two implementations of our method using logistic regressions and gradient boosted trees, together with a simple training procedure based on Expectation Maximization. We further derive an efficient prediction procedure based on dynamic programming, thus avoiding the cost of examining an exponential number of potential label subsets. Experimental results show the effectiveness of the proposed method against competitive alternatives on benchmark datasets.
RIS
TY - CPAPER TI - Conditional Bernoulli Mixtures for Multi-label Classification AU - Cheng Li AU - Bingyu Wang AU - Virgil Pavlu AU - Javed Aslam BT - Proceedings of The 33rd International Conference on Machine Learning PY - 2016/06/11 DA - 2016/06/11 ED - Maria Florina Balcan ED - Kilian Q. Weinberger ID - pmlr-v48-lij16 PB - PMLR SP - 2482 DP - PMLR EP - 2491 L1 - http://proceedings.mlr.press/v48/lij16.pdf UR - http://proceedings.mlr.press/v48/lij16.html AB - Multi-label classification is an important machine learning task wherein one assigns a subset of candidate labels to an object. In this paper, we propose a new multi-label classification method based on Conditional Bernoulli Mixtures. Our proposed method has several attractive properties: it captures label dependencies; it reduces the multi-label problem to several standard binary and multi-class problems; it subsumes the classic independent binary prediction and power-set subset prediction methods as special cases; and it exhibits accuracy and/or computational complexity advantages over existing approaches. We demonstrate two implementations of our method using logistic regressions and gradient boosted trees, together with a simple training procedure based on Expectation Maximization. We further derive an efficient prediction procedure based on dynamic programming, thus avoiding the cost of examining an exponential number of potential label subsets. Experimental results show the effectiveness of the proposed method against competitive alternatives on benchmark datasets. ER -
APA
Li, C., Wang, B., Pavlu, V. & Aslam, J.. (2016). Conditional Bernoulli Mixtures for Multi-label Classification. Proceedings of The 33rd International Conference on Machine Learning, in PMLR 48:2482-2491

Related Material