Meta Path-based Recommendation System On Heterogeneous Information Network

With the evolution of computer science and Internet,social netwok anal-ysis becomes an important task for data mining.Lots of works on social net-wok analysis focus on homogeneous information network.But with the data information in Internet becoming more and more various,research on hetero-geneous information network(HIN)has attracted a lot of attention.Many data mining tasks,such as classification,clustering,ranking,similarity measure and so on,utitlze HIN to integrate different kinds of heterogeneous information and achieve significant improvements.As an important data mining task,recom-mendation can help users to find their favorite items and alleviates information overload problem in the Internet.A real-world recommender system includes a lot of object types(e.g.,users,movies,actors,and interest groups in movie recommendation)and the rich relations among object types,which naturally constitute a HIN.However,conventional HINs do not consider the attribute values on links,and the widely used meta path in HIN may fail to accurately capture semantic relations among objects,due to the existence of rating scores(usually ranging from 1 to 5)be-tween users and items in recommender system.Otherwise,how to utilize HIN to integrate multiple information and learn personalized and prioritized pref-erence weight for each user is another key problem of this paper.Aims at the problems above,this paper contributes mainly in three apsects.Firstly,it introduces the weighted HIN and weighted meta path concepts to subtly depict the path semantics through distinguishing different link attribute values.It also proposes a strategy for similarity measure along weighted meta path,which alows the existing similarity measures to apply in weighted meta path after being modified.Secondly,it proposes a semantic path based personalized recommendation method SemRec to predict the rating scores of users on items.Through setting meta paths,SemRec not only flexibly integrates heterogeneous information but also obtains prioritized and personalized weights representing user preferences on paths.Experiments on three real datasets illustrate that SemRec achieves better recommendation performance through flexibly integrating information with the help of weighted meta paths.Moreover,extensive experiments validate the benefits of weighted meta paths.At last,based on the proposal SemRec,a movie recommender system is de-signed for explaining recommended results according to the personalized pref-erence weights learned.