A User-Centric Diversity by Design Recommender System for the Movie Application Domain

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

Abstract

Recommender systems (RS) have seen widespread adoption across the Internet. However, by emphasizing personalization through the optimization of accuracy-focused metrics, over-personalization may emerge, with negative effects on the user experience. A countermeasure to the problem is to diversify recommendations. In this paper, we present a solution that addresses the problem in the context of a movie application domain. The solution enhances diversity on four related dimensions, namely global coverage, local coverage, novelty, and redundancy. The proposed solution is designed to diversify users profiles, modeled on categorical preferences, within the same group in the recommendation filtering. We evaluate our approach on the Movielens dataset and show that our algorithm yields better results compared to random selection distant neighbors and performs comparably to one of the current state of the art solutions.
Close

Details

Recommender systems (RS) have seen widespread adoption across the Internet. However, by emphasizing personalization through the optimization of accuracy-focused metrics, over-personalization may emerge, with negative effects on the user experience. A countermeasure to the problem is to diversify recommendations. In this paper, we present a solution that addresses the problem in the context of a movie application domain. The solution enhances diversity on four related dimensions, namely global coverage, local coverage, novelty, and redundancy. The proposed solution is designed to diversify users profiles, modeled on categorical preferences, within the same group in the recommendation filtering. We evaluate our approach on the Movielens dataset and show that our algorithm yields better results compared to random selection distant neighbors and performs comparably to one of the current state of the art solutions.
Original languageEnglish
Title of host publicationWWW '18 Companion Proceedings of the The Web Conference 2018
Number of pages9
PublisherAssociation for Computing Machinery
Publication date23 Apr 2018
Pages1381-1389
ISBN (Electronic)978-1-4503-5640-4
DOI
Publication statusPublished - 23 Apr 2018
Publication categoryResearch
Peer-reviewedYes
EventThe Web Conference 2018 - Workshop on Online Recommender Systems and User Modeling: continuous learning from web data - Lyon, France
Duration: 23 Apr 201827 Apr 2018
https://www2018.thewebconf.org

Conference

ConferenceThe Web Conference 2018 - Workshop on Online Recommender Systems and User Modeling: continuous learning from web data
LocationLyon
LandFrance
Periode23/04/201827/04/2018
Internetadresse

    Research areas

  • diversity, Recommender System, movie recommendation, user clustering, diversification

Map

Download statistics

No data available
ID: 271454152