Individual Persistence Adaptation for User-Centric Evaluation of User Satisfaction in Recommender Systems
Nozomu Onodera,Keisuke Maeda,Takahiro Ogawa,Miki Haseyama
DOI: https://doi.org/10.1109/access.2024.3360693
IF: 3.9
2024-02-20
IEEE Access
Abstract:The primary objective of a recommender system (RS) is to enhance user satisfaction, which serves as the gold standard for evaluation. In order to support the advancement of RS, it is crucial to study how to accurately measure user satisfaction. This paper proposes a novel evaluation framework that leverages user-centric evaluations to achieve a more precise measurement of satisfaction. User-centric evaluations involve metrics that model the user's behavior toward the recommended ranking list. These metrics incorporate a persistence parameter that captures the user's personality traits. However, most conventional studies in the RS field utilize a predefined persistence parameter for all users, neglecting the individual differences among users. In contrast, our work focuses on optimizing this parameter individually for each user, which, to the best of our knowledge, is the first attempt in the RS context. To demonstrate the effectiveness of our framework in measuring satisfaction, we conducted innovative subject experiments to collect satisfaction data. Through these experiments, we obtained valuable insights into user satisfaction, which we used for quantitative correlation analysis. The results of this analysis provide empirical evidence that our framework is well-suited for RS evaluation, showing a stronger alignment with user satisfaction compared to conventional approaches.
computer science, information systems,telecommunications,engineering, electrical & electronic