A Dichotomous Group Recommendation Algorithm Based on User Fairness
Hongyao Zhong, Junjie Jia · 2023
Group recommendation systems that target users with several objectives have gained more attention in recent years because of the quick development of computers. To include group preferences, group recommendation algorithm systems currently employ a weighted technique. This tactic, however, ignores the variations in preferences among group members, escalating confrontations over preferences. This research suggests a user divergence-based binary group recommendation method to overcome this problem. The algorithm creates a top-k recommendation list for each user using the user recommendation system, then merges the recommendation lists inside the group. A randomized greedy algorithm is used on the combined list in each iteration to produce a top-2k list. To generate the final group recommendation list, top-k, the algorithm traverses the top-2k list while minimizing the greatest divergence. In this study, user divergence is defined, and its boundedness, monotonicity, and optimality are demonstrated. The suggested method displays great recommendation accuracy and fairness even under diverse group sizes and random group recommendation lists, according to experimental data.