Design And Implementation Of Recommendation System Based On Event-based Social Network

An event-based social network(EBSN)has emerged as human beings step into the information age,and it has developed rapidly in recent years.And it is difficult for users to quickly obtain the content they need because of increasing social network information,there has been increasing attention to the issue of activity recommendations in EBSN.Although relevant research is increasing,there are special difficulties in making personalized recommendations for activities in EBSN:the number of users participating in activities is very limited,and the number of participated users in a single activity is also limited,and some of the activities are one-off activities,so The user-activity matrix is very sparse;the social and offline nature of EBSN determines that the activity recommendation needs to meet the user’s online and offline social relationships and personal behavior patterns at the same time,and traditional recommendation methods cannot achieve better results.And with the development of EBSN,the corresponding recommended methods need to be updated synchronously and iteratively,which will cause frequent system upgrades.In order to solve these problems,this paper proposes an improved collaborative filtering recommendation method that integrates multiple feature factors,and uses microservice technology to construct a recommendation framework,designs and implements a recommendation system based on active social networks.The main contents of the paper are as follows:(1)Design a recommendation framework based on micro-services.Personalized recommendation With the development of the system and the accumulation of data,the recommendation model itself also needs to be updated iteratively.The system implemented by traditional web technology needs to constantly update the entire system during iteration.This article uses micro-service technology to design a recommendation framework.Micro-services implement different recommendation recall strategies and ranking strategies.When iterative update of the recommendation model is required,only the corresponding micro-services need to be updated to complete the iterative upgrade of the entire system function.The micro-service-based recommendation framework also provides the possibility of flexible combination of multiple strategies,enabling the system to customize the system recommendation strategy according to the usage scenario and find the most suitable recommendation method in a certain scenario.(2)Design an improved collaborative filtering recommendation model that integrates multiple feature factors.Existing social network recommendation methods for activities mostly start from factors that affect users’ participation in activities,or start recommendations based on a graph model constructed based on users’ social relationships.Based on the analysis of user behavior characteristics,information characteristics,activity characteristics,social relations,publisher influence and other characteristic data,combined with the collaborative filtering model,this paper combines user behavior patterns and social relations and other influencing factors with the collaborative filtering model to construct a Fusion recommendation model.In the end,it will provide users with higher satisfaction recommendation results.(3)Design and implement a recommendation system based on active social networks.In this paper,the system is designed into two subsystems,namely the business subsystem that uses web development technology to implement business processes,and the recommendation subsystem that uses micro-service technology to implement.Data sharing and interaction between the two subsystems realizes the basic functions of users using the ENSN activity recommendation system.Finally,the function and performance of the system are tested.