Clustering-based recommender system: Bundle recommendation using matrix factorization to single user and user communities

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper shows the results of a Recommender System (RS) that suggests bundles of items to a user or a community of users. Nowadays, there are several RS that realize suggestions of a unique item considering the preferences of a user. However, these RS are not scalable and sometimes the suggestions that make are far from a user’s preferences. We propose an RS that suggests bundles of items to one user or a community of users with similar affinities. This RS uses an algorithm based on Matrix Factorization (MF). To execute the experiments, we use released databases with high dispersion. The results obtained are evaluated per the metrics Accuracy, Precision, Recall and F-measure. The results demonstrate that the proposed method improves significantly the quality of the suggestions.

Original languageEnglish
Title of host publicationClustering-based recommender system: Bundle recommendation using matrix factorization to single user and user communities
EditorsTareq Z. Ahram
Pages330-338
Number of pages9
ISBN (Electronic)9783319942285
DOIs
StatePublished - 1 Jan 2019
EventAdvances in Intelligent Systems and Computing - , Germany
Duration: 1 Jan 2015 → …

Publication series

NameAdvances in Intelligent Systems and Computing
Volume787
ISSN (Print)2194-5357

Conference

ConferenceAdvances in Intelligent Systems and Computing
Country/TerritoryGermany
Period1/01/15 → …

Keywords

  • Bundles of items
  • Matrix factorization
  • Recommender system

Fingerprint

Dive into the research topics of 'Clustering-based recommender system: Bundle recommendation using matrix factorization to single user and user communities'. Together they form a unique fingerprint.

Cite this