The Design And Implementation Of Event Recommender System Based On Preference Migration And Distance Perception

With the continuous development of social networks,a novel heterogeneous social network has emerged:Event Based Social Network(EBSN).In order to deal with the problem of "information overload" in EBSN,the research of event recommender system has been put on the agenda.Unlike other types of recommender systems,the characteristics of EBSN make the event recommender system face many problems:Due to the inherent ordering of the event itself,the preferences of group participation events will continuously migrate;Different users have different sensitivity to event distance,which greatly affects users' event participation.In order to solve the above problems,this article designs and implements an event recommender system.The main research contents are as follows:(1)Aiming at the problem that most existing event recommendation models ignore the inherent potential relationship between events,a group event recommendation model based on preference migration is proposed.This model makes full use of the user's historical event participation information to model the extracted context information such as user preferences and group topics,and then uses the gibbs sampling algorithm to generate the topic model,and then calculate the preference migration coefficient of each event.Finally combined with the recommendation model to calculate the score of each candidate event and rank the candidate event to complete the Top-N recommendation.(2)Aiming at the problem of insufficient utilization of event location information by the existing event recommendation models,and combining the sensitivity of users to different event distances,a group event recommendation model based on distance perception was proposed.The model uses a clustering method to cluster events into several regions according to latitude and longitude information,and uses the sensitivity of each user to the event distance in the user's event participation record as a feature,and combines the other context information of the event to model.Then the event score is obtained by calculating the cosine similarity of model features of each candidate event,so as to complete the Top-N recommendation.(3)An event recommendation model based on preference migration and distance perception is proposed.The region information of event is added to the topic model,and the model is trained by the gibbs sampling algorithm.Then the inherent relationship between events is used to influence the recommendation model and calculate event scores,thus improving the recommendation accuracy.Experimental tests on public data sets show that the algorithm achieves better recommendation results.(4)Based on the model proposed above,an event recommender system based on preference migration and distance perception is designed and implemented.This system can fully explore the inherent potential relationship between events,fully meet the needs of users' participation in events,and effectively present the recommendation results to users.