HRFA: Don’t Ignore Strangers with Different Views

Senhui Zhang, Wendi Ji, Jiahao Yuan, Xiaoling Wang

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

Abstract

Review-based recommender suffers from the sparsity of reviews: only a few users leave substantial comments in the real world. As a result, some recent methods resort to supplementary reviews written by similar users, which only leverage homogeneous preferences. However, users holding different views could also supply valuable information with heterogeneous preferences. In this paper, we propose a recommendation model for rating prediction, named Heterogeneous Review-based Recommendation via Four-way Attention (HRFA). To take advantage of the heterogeneous preferences, the supplementary reviews in HRFA are redefined as reviews from all users with common purchase history, no matter whether they give similar ratings. Specially, we integrate the heterogeneous preferences into the one semantic space via introducing a similarity projection based on rating difference. Experiments conducted on five datasets demonstrate that our model achieves higher rating prediction accuracy than other baselines.

Original languageEnglish
Title of host publicationWeb Information Systems Engineering - WISE 2021 - 22nd International Conference on Web Information Systems Engineering, WISE 2021, Proceedings
EditorsWenjie Zhang, Lei Zou, Zakaria Maamar, Lu Chen
PublisherSpringer Science and Business Media Deutschland GmbH
Pages209-217
Number of pages9
ISBN (Print)9783030915599
DOIs
StatePublished - 2021
Event22nd International Conference on Web Information Systems Engineering, WISE 2021 - Melbourne, Australia
Duration: 26 Oct 202129 Oct 2021

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13081 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference22nd International Conference on Web Information Systems Engineering, WISE 2021
Country/TerritoryAustralia
CityMelbourne
Period26/10/2129/10/21

Keywords

  • Data sparsity
  • Heterogeneous preferences
  • Rating prediction
  • Review-based recommendation

Fingerprint

Dive into the research topics of 'HRFA: Don’t Ignore Strangers with Different Views'. Together they form a unique fingerprint.

Cite this