Research On Heterogeneous Social Network Entity Anchor Link Recognition Technology Based On Meta-path

With the rapid development of network technology,users can use multiple online social networks to enjoy services in their daily life.At the same time,there are many kinds of social network service platforms,which gradually form a huge social network structure.On the whole,most social networks are heterogeneous social networks,and the entities in heterogeneous social networks mainly include users,location,text and so on.Heterogeneous social network entity anchor link recognition can promote the development of social network and continuously improve the social service platform,which is an important research content in social network.This thesis mainly focuses on the anchor link recognition technology of users and locations in heterogeneous social networks.Firstly,aiming at the problem of anchor link recognition of heterogeneous social network users,this thesis proposes an anchor link recognition algorithm based on meta-path based on the link relationship and attribute information of users.In addition to the direct link relationship between users and inherent attributes,user entities also have the indirect link relationship between users and indirect attributes generated by social activities.In this thesis,the meta-path technique is used to fully mine the indirect link relations and indirect attributes generated by users' social activities,compare the pattern consistency of user link relations between different networks,and use different similarity calculation methods to measure the similarity of attribute information between users.The experimental results show that the proposed meta-path based anchor link recognition algorithm has higher accuracy.Secondly,aiming at the problem of location anchor link recognition in heterogeneous social networks,this thesis proposes a location anchor link recognition algorithm based on meta-path.At present,there is little research on anchor link recognition of positional entity.Different from user entities,location entities do not have subjective initiative.In heterogeneous social networks,the link relationship between location entities is relatively fixed,and its own attribute information is not rich enough,so it is difficult to identify location anchor link.In this thesis,based on the prior knowledge of a small number of anchor link users,the relationship between the location checked in by users is mined by meta-path technology.On this basis,the location anchor link recognition model is established by combining the location attribute information.Experiments show that the proposed method based on meta-path has high accuracy.