A fast algorithm for learning large scale preference relations

Vikas C. Raykar, Ramani Duraiswami, Balaji Krishnapuram
Proceedings of the Eleventh International Conference on Artificial Intelligence and Statistics, PMLR 2:388-395, 2007.

Abstract

We consider the problem of learning the ranking function that maximizes a generalization of the Wilcoxon-Mann-Whitney statistic on training data. Relying on an -exact approximation for the error-function, we reduce the computational complexity of each iteration of a conjugate gradient algorithm for learning ranking functions from O(m^2), to O(m), where m is the size of the training data. Experiments on public benchmarks for ordinal regression and collaborative filtering show that the proposed algorithm is as accurate as the best available methods in terms of ranking accuracy, when trained on the same data, and is several orders of magnitude faster.

Cite this Paper

BibTeX
@InProceedings{pmlr-v2-raykar07a, title = {A fast algorithm for learning large scale preference relations}, author = {Raykar, Vikas C. and Duraiswami, Ramani and Krishnapuram, Balaji}, booktitle = {Proceedings of the Eleventh International Conference on Artificial Intelligence and Statistics}, pages = {388--395}, year = {2007}, editor = {Meila, Marina and Shen, Xiaotong}, volume = {2}, series = {Proceedings of Machine Learning Research}, address = {San Juan, Puerto Rico}, month = {21--24 Mar}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v2/raykar07a/raykar07a.pdf}, url = {https://proceedings.mlr.press/v2/raykar07a.html}, abstract = {We consider the problem of learning the ranking function that maximizes a generalization of the Wilcoxon-Mann-Whitney statistic on training data. Relying on an -exact approximation for the error-function, we reduce the computational complexity of each iteration of a conjugate gradient algorithm for learning ranking functions from O(m^2), to O(m), where m is the size of the training data. Experiments on public benchmarks for ordinal regression and collaborative filtering show that the proposed algorithm is as accurate as the best available methods in terms of ranking accuracy, when trained on the same data, and is several orders of magnitude faster.} }
Endnote
%0 Conference Paper %T A fast algorithm for learning large scale preference relations %A Vikas C. Raykar %A Ramani Duraiswami %A Balaji Krishnapuram %B Proceedings of the Eleventh International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2007 %E Marina Meila %E Xiaotong Shen %F pmlr-v2-raykar07a %I PMLR %P 388--395 %U https://proceedings.mlr.press/v2/raykar07a.html %V 2 %X We consider the problem of learning the ranking function that maximizes a generalization of the Wilcoxon-Mann-Whitney statistic on training data. Relying on an -exact approximation for the error-function, we reduce the computational complexity of each iteration of a conjugate gradient algorithm for learning ranking functions from O(m^2), to O(m), where m is the size of the training data. Experiments on public benchmarks for ordinal regression and collaborative filtering show that the proposed algorithm is as accurate as the best available methods in terms of ranking accuracy, when trained on the same data, and is several orders of magnitude faster.
RIS
TY - CPAPER TI - A fast algorithm for learning large scale preference relations AU - Vikas C. Raykar AU - Ramani Duraiswami AU - Balaji Krishnapuram BT - Proceedings of the Eleventh International Conference on Artificial Intelligence and Statistics DA - 2007/03/11 ED - Marina Meila ED - Xiaotong Shen ID - pmlr-v2-raykar07a PB - PMLR DP - Proceedings of Machine Learning Research VL - 2 SP - 388 EP - 395 L1 - http://proceedings.mlr.press/v2/raykar07a/raykar07a.pdf UR - https://proceedings.mlr.press/v2/raykar07a.html AB - We consider the problem of learning the ranking function that maximizes a generalization of the Wilcoxon-Mann-Whitney statistic on training data. Relying on an -exact approximation for the error-function, we reduce the computational complexity of each iteration of a conjugate gradient algorithm for learning ranking functions from O(m^2), to O(m), where m is the size of the training data. Experiments on public benchmarks for ordinal regression and collaborative filtering show that the proposed algorithm is as accurate as the best available methods in terms of ranking accuracy, when trained on the same data, and is several orders of magnitude faster. ER -
APA
Raykar, V.C., Duraiswami, R. & Krishnapuram, B.. (2007). A fast algorithm for learning large scale preference relations. Proceedings of the Eleventh International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 2:388-395 Available from https://proceedings.mlr.press/v2/raykar07a.html.

Related Material