Dynamic modeling of user preferences for stable recommendations

Oluwafemi Olaleke, Ivan Oseledets, Evgeny Frolov

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

2 Citations (Scopus)

Abstract

In domains where users tend to develop long-term preferences that do not change too frequently, the stability of recommendations is an important factor of the perceived quality of a recommender system. In such cases, unstable recommendations may lead to poor personalization experience and distrust, driving users away from a recommendation service. We propose an incremental learning scheme that mitigates such problems via the dynamic modeling approach. It incorporates a generalized matrix form of a partial differential equation integrator that yields a dynamic low-rank approximation of time-dependent matrices representing user preferences. The scheme allows extending the famous PureSVD approach to time-aware settings and significantly improves its stability without sacrificing the accuracy in standard top-n recommendations tasks.

Original languageEnglish
Title of host publicationUMAP 2021 - Proceedings of the 29th ACM Conference on User Modeling, Adaptation and Personalization
PublisherAssociation for Computing Machinery, Inc
Pages262-266
Number of pages5
ISBN (Electronic)9781450383660
DOIs
Publication statusPublished - 21 Jun 2021
Event29th ACM Conference on User Modeling, Adaptation and Personalization, UMAP 2021 - Virtual, Online, Netherlands
Duration: 21 Jun 202025 Jun 2020

Publication series

NameUMAP 2021 - Proceedings of the 29th ACM Conference on User Modeling, Adaptation and Personalization

Conference

Conference29th ACM Conference on User Modeling, Adaptation and Personalization, UMAP 2021
Country/TerritoryNetherlands
CityVirtual, Online
Period21/06/2025/06/20

Keywords

  • Collaborative Filtering
  • PureSVD
  • Recommendations Stability
  • User Preferences Dynamics

Fingerprint

Dive into the research topics of 'Dynamic modeling of user preferences for stable recommendations'. Together they form a unique fingerprint.

Cite this