Less Can Be More: Exploring Population Rating Dispositions with Partitioned Models in Recommender Systems
Open Access
- 16 June 2023
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 291-295
Abstract
In this study, we partition users by rating disposition - looking first at their percentage of negative ratings, and then at the general use of the rating scale. We hypothesize that users with different rating dispositions may use the recommender system differently and therefore the agreement with their past ratings may be less predictive of the future agreement. We use data from a large movie rating website to explore whether users should be grouped by disposition, focusing on identifying their various rating distributions that may hurt recommender effectiveness. We find that such partitioning not only improves computational efficiency but also improves top-k performance and predictive accuracy. Though such effects are largest for the user-based KNN CF, smaller for item-based KNN CF, and smallest for latent factor algorithms such as SVD.Keywords
All Related Versions
This publication has 8 references indexed in Scilit:
- Surprise: A Python library for recommender systemsJournal of Open Source Software, 2020
- Why are online reviews in the sharing economy skewed toward positive ratings? Linking customer perceptions of service quality to leaving a review of an Airbnb stayJournal of Tourism Futures, 2020
- How scales influence user rating behaviour in recommender systemsBehaviour & Information Technology, 2017
- Improving Collaborative Filtering Based Recommenders Using Topic Modelling2014
- Rating support interfaces to improve user experience and recommender accuracy2013
- Users and Noise: The Magic Barrier of Recommender Systems2012
- The Impact of Rating Scales on User’s Rating Behavior2011
- Overcoming the J-shaped distribution of product reviewsCommunications of the ACM, 2009