Group Recommendation Method Based On Co-evolution Of Group Preference And User Preference

The group recommendation system has become an important tool of social platforms to provide personalized and satisfied products or services for groups.Most of the existing group recommendation methods are the integration and aggregation of personalized recommendation methods,but they ignore the dynamic changes preferences and the interaction between groups and users,which cannot guarantee the effectiveness and performance of group recommendation systems.Therefore,in order to accurately model the dynamic evolution of group preferences,this paper has done the following related work based on relevant theories and research.(1)Group Recommendation Model Based on Group Preference Evolution.When modeling group preferences,considering the dynamics of group preferences and the dependence of group preferences on historical preferences,it is believed that the evolution of group preferences is related to the group's previous preferences.The experimental results on the Deviant Art dataset show that considering the time factor when modeling group preferences can improve the accuracy of group recommendations.(2)Group Recommendation Model Based on Both Group and User Preference Evolution.When modeling group preferences,considering the changes of group members,it is considered that group preferences are affected by both group historical preferences and the preferences of new members in the group.The experimental results on the Deviant Art dataset show that group recommendation can be improved by considering group member changes when modeling group preferences.(3)Group Recommendation Model Based on Co-evolution of Group preference and User preference.When modeling the evolution of group preferences,both the historical dependence of group preferences and the changes of group members,as well as the dynamic interaction and influence of groups and users,are considered.The group preferences and user preferences interact with each other during the evolution process.The experimental results on the Deviant Art dataset show the effectiveness of predicting group preferences by considering the interaction between groups and users.