Relational Collaborative Filtering: Modeling Multiple Item Relations for Recommendation
- Submitting institution
-
University of Glasgow
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 11-12081
- Type
- E - Conference contribution
- DOI
-
10.1145/3331184.3331188
- Title of conference / published proceedings
- 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR'19)
- First page
- 125
- Volume
- -
- Issue
- -
- ISSN
- -
- Open access status
- -
- Month of publication
- -
- Year of publication
- 2019
- URL
-
http://eprints.gla.ac.uk/191181/
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
4
- Research group(s)
-
-
- Citation count
- 8
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- ORIGINALITY: We have developed a novel hierarchical self-attention based model for capturing fine grained item-relationships from multiple perspectives. RIGOUR: This paper is published in the top Information Retrieval conference (20% acceptance rate). This work has a mathematical basis and is rigorously evaluated on two real-life data sets (MovieLens and WSDM Cup Challenge data) against seven state-of-the-art baselines. SIGNIFICANCE: We demonstrate significant improvements, ranging 11-29%, on the state-of-the-art baselines. We have shown the importance of relation type and value for collaborative filtering and highly applicable in real-life recommendation applications (e.g., Netflix, Amazon), where fine grained relationships improve effectiveness.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -