Toward a Cognitive Model for Group Recommender Systems

Abstract

In the contemporary world, the abundance of data and the growing number of users in virtual environments have led to a focus on classifying data and grouping users to create recommendation lists that enhance user experiences. While existing research has explored various methods in developing group recommender systems, the role of incoming flows and their impacts in the decision-making process of joining a group has been largely overlooked. This study, drawing on cognitive theories, aims to examine user behavior in decision-making and group assignment. Through the mapping of users' interactions in complex dynamic networks, users are categorized based on their incoming interactions. A key feature of this study is its consideration of users belonging to multiple groups simultaneously, with each group’s behavioral profile being delineated accordingly. The proposed method simulates real-world decision-making processes to provide tailored recommendations for each individual. The effectiveness of our approach is assessed using four datasets.