Rank Aggregation for Presentation-Centric Preference Estimation

Evgeny Televitckiy, Carmel Domshlak

We consider the connection between the recommendations' selection process and the form in which these recommendations are then presented to the users. On the one hand, a user of an e-commerce site typically provides her preferences as a (very rough) weak ordering over a small subset of all the available items. On the other hand, the recommendations selected for a user are presented to her as this or another projection of a strict ordering over the whole space of items. Aiming at bridging the gap between these two properties of the systems, we suggest and evaluate techniques for rank aggregation in this setting.

Subjects: 12. Machine Learning and Discovery; Please choose a second document classification

Submitted: May 12, 2007

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.