Preference-Based Rank Elicitation using Statistical Models: The Case of Mallows

Robert Busa-Fekete, Eyke Huellermeier, Balázs Szörényi
Proceedings of the 31st International Conference on Machine Learning, PMLR 32(2):1071-1079, 2014.

Abstract

We address the problem of rank elicitation assuming that the underlying data generating process is characterized by a probability distribution on the set of all rankings (total orders) of a given set of items. Instead of asking for complete rankings, however, our learner is only allowed to query pairwise preferences. Using information of that kind, the goal of the learner is to reliably predict properties of the distribution, such as the most probable top-item, the most probable ranking, or the distribution itself. More specifically, learning is done in an online manner, and the goal is to minimize sample complexity while guaranteeing a certain level of confidence.

Cite this Paper

BibTeX
@InProceedings{pmlr-v32-busa-fekete14, title = {Preference-Based Rank Elicitation using Statistical Models: The Case of Mallows}, author = {Busa-Fekete, Robert and Huellermeier, Eyke and Szörényi, Balázs}, booktitle = {Proceedings of the 31st International Conference on Machine Learning}, pages = {1071--1079}, year = {2014}, editor = {Xing, Eric P. and Jebara, Tony}, volume = {32}, number = {2}, series = {Proceedings of Machine Learning Research}, address = {Bejing, China}, month = {22--24 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v32/busa-fekete14.pdf}, url = {https://proceedings.mlr.press/v32/busa-fekete14.html}, abstract = {We address the problem of rank elicitation assuming that the underlying data generating process is characterized by a probability distribution on the set of all rankings (total orders) of a given set of items. Instead of asking for complete rankings, however, our learner is only allowed to query pairwise preferences. Using information of that kind, the goal of the learner is to reliably predict properties of the distribution, such as the most probable top-item, the most probable ranking, or the distribution itself. More specifically, learning is done in an online manner, and the goal is to minimize sample complexity while guaranteeing a certain level of confidence.} }
Endnote
%0 Conference Paper %T Preference-Based Rank Elicitation using Statistical Models: The Case of Mallows %A Robert Busa-Fekete %A Eyke Huellermeier %A Balázs Szörényi %B Proceedings of the 31st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2014 %E Eric P. Xing %E Tony Jebara %F pmlr-v32-busa-fekete14 %I PMLR %P 1071--1079 %U https://proceedings.mlr.press/v32/busa-fekete14.html %V 32 %N 2 %X We address the problem of rank elicitation assuming that the underlying data generating process is characterized by a probability distribution on the set of all rankings (total orders) of a given set of items. Instead of asking for complete rankings, however, our learner is only allowed to query pairwise preferences. Using information of that kind, the goal of the learner is to reliably predict properties of the distribution, such as the most probable top-item, the most probable ranking, or the distribution itself. More specifically, learning is done in an online manner, and the goal is to minimize sample complexity while guaranteeing a certain level of confidence.
RIS
TY - CPAPER TI - Preference-Based Rank Elicitation using Statistical Models: The Case of Mallows AU - Robert Busa-Fekete AU - Eyke Huellermeier AU - Balázs Szörényi BT - Proceedings of the 31st International Conference on Machine Learning DA - 2014/06/18 ED - Eric P. Xing ED - Tony Jebara ID - pmlr-v32-busa-fekete14 PB - PMLR DP - Proceedings of Machine Learning Research VL - 32 IS - 2 SP - 1071 EP - 1079 L1 - http://proceedings.mlr.press/v32/busa-fekete14.pdf UR - https://proceedings.mlr.press/v32/busa-fekete14.html AB - We address the problem of rank elicitation assuming that the underlying data generating process is characterized by a probability distribution on the set of all rankings (total orders) of a given set of items. Instead of asking for complete rankings, however, our learner is only allowed to query pairwise preferences. Using information of that kind, the goal of the learner is to reliably predict properties of the distribution, such as the most probable top-item, the most probable ranking, or the distribution itself. More specifically, learning is done in an online manner, and the goal is to minimize sample complexity while guaranteeing a certain level of confidence. ER -
APA
Busa-Fekete, R., Huellermeier, E. & Szörényi, B.. (2014). Preference-Based Rank Elicitation using Statistical Models: The Case of Mallows. Proceedings of the 31st International Conference on Machine Learning, in Proceedings of Machine Learning Research 32(2):1071-1079 Available from https://proceedings.mlr.press/v32/busa-fekete14.html.

Related Material