Skip to main navigation Skip to search Skip to main content

Optimal recommender systems blending

  • Laboratoire d'Informatique (LIX)

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In the Recommender Systems field ensemble techniques gain growing interest. This approach is based on the idea of mixing many recommenders and to get an average prediction from all of them. Even if it is useful this process may be very expensive from a computational point of view. We propose the use of Operations Research techniques in order to optimize the balance of different predictors and to accelerate it. We show that this problem can be generalized, thus we provide a mathematical framework which helps to find further improvements.

Original languageEnglish
Title of host publicationWIMS'11 - Proceedings of the International Conference on Web Intelligence, Mining and Semantics
PublisherAssociation for Computing Machinery
ISBN (Print)9781450301480
DOIs
Publication statusPublished - 1 Jan 2011

Publication series

NameACM International Conference Proceeding Series

Keywords

  • collaborative filtering
  • optimization
  • recommender systems

Fingerprint

Dive into the research topics of 'Optimal recommender systems blending'. Together they form a unique fingerprint.

Cite this