A dynamic fuzzy group recommender system based on intuitionistic fuzzy choquet integral aggregation

In this study, a new dynamic fuzzy group recommender system (DFGRS) is investigated, which aims to deal with some critical challenges in the group recommender system (GRS), including the user preference uncertainty, item attractiveness tendency, and group members’ interaction and fairness. The proposed approach of the DFGRS applies intuitionistic fuzzy sets (IFSs) to represent user preferences on the considered items by taking into account user hesitancy and inconsistency in ratings. It also incorporates a novel dynamic similarity mechanism for taking into account users’ preference tendency and item attractiveness in predicting unknown user-item preferences. The model’s parameters are optimized using a modified Bayesian method. This combination creates a more realistic and concise prediction of unknown user ratings. Then, to form a balanced solution in the consensus phase of the DFGRS, an intuitionistic fuzzy Choquet integral aggregation operation (IFCAO) is suggested. This approach enables the DFGRS to generate recommendations that balance between optimal average group users’ preferences and fairness among them. Furthermore, the proposed IFCAO includes the group members’ interaction impact due to the proposed capacity function. In addition, several other intuitionistic fuzzy aggregation operations can be utilized that inherit features of common aggregation strategies defined for a traditional GRS. The experimental results in this paper demonstrate the performance of various aggregation approaches for the consensus phase of DFGRS. The obtained performance analysis results highlight the advantage and the potential of the DFGRS based on the proposed Choquet integral aggregation operation. Overall, the proposed DFGRS presents a novel approach for addressing challenges in group recommendation and could have important practical applications in various areas.