Research On Multi-dimensional Social Network Link Prediction Based On Network Embedding

With the rapid development of the Internet,the social network has become a part of people's daily life.The interactive behaviors among users have produced a large amount of interactive data,which reflects some common phenomena of social development.The analysis and mining of user interaction data has very important research significance and application value for the development of the Internet and related industries.This article starts from two aspects of multi-dimensional social network attributes and network embedded representation,and studies multi-dimensional social network link prediction methods.The research contents are:(1)Multi-dimensional social network representation based on network embedding.Finding a meaningful network representation is a prerequisite for effective mining and analysis of network data.Traditional theory represents the network as a high-dimensional and sparse adjacency matrix;while network embedding represents the network as a low-dimensional space where each node is represented by a low-dimensional vector.The vector representation embedded in the network can not only maintain the proximity characteristics between nodes,but also can be used as practical input for learning input,node classification and link prediction.By learning the representation method of multi-dimensional networks,you can capture the independent information of each dimension and the dependent information of each dimension.Therefore,based on the analysis of multi-dimensional social networks,a multi-dimensional network embedded representation framework is proposed.The node2 vec method is used to learn the continuous feature representation of nodes in the network.The network neighborhood of the nodes is preserved to the greatest extent.Representation in a coherent model of learning.(2)Multi-dimensional social network link prediction based on attention model.The purpose of learning representation in multi-dimensional networks is to infer robust node representations between views of different dimensions,and to promote collaboration among different views.Based on multi-dimensional network embedding,a social network link prediction method based on attention model is proposed.In the process of view information integration,assign appropriate weights to views,so that each node is concentrated on the view with the most information,so as to select a robust node representation,realize link prediction and improve accuracy.(3)Combined with real DBLP,PPI,Flickr,You Tube,and Twitter social network datasets,the commonly used evaluation indicators were used to verify the model through node classification and link prediction tasks.Three sets of data sets are used for node classification,and the other two sets are used for link prediction.The experimental results show that the method proposed in this study is feasible and effective.Multi-dimensional social network link prediction based on network embedding provides a new research idea for social network link prediction research based on analyzing the embedded representation of social network and combining attention mechanism.