The Research Of Link Prediction Based On Closeness And Node Contribution

With the rapid development of Internet technology,social networks gradually comes into the people's vision in recent years.Scholars' discussion of social networks is becoming more and more intense.Link prediction,as one of the social network research problems,not only has great value in the analysis of network evolution,but also plays an important role in practical application.In this paper,the problem of link prediction in social network is studied.The main contents are as follows.Firstly,the research status of social network link forecasting is introduced,and the relationship between graph theory and network,network storage form and network attribute are introduced and analyzed.The network attribute plays an important role in the prediction performance of network.It also describes the link prediction definition,studies the problem and summarizes the different categories based on the similarity link prediction algorithm.Secondly,a closeness index algorithm based on the local information algorithm is proposed.This algorithm proposes a concept of closeness for CN algorithm,only considering common neighbor nodes' ignoring the relationship between them.The algorithm considers the degree of correlation of network nodes,and the relationship between the predicted nodes and the relationships among the small groups within the common neighbor nodes are quantified by two different closeness.A new node similarity algorithm is proposed by closeness to achieve the prediction of the link.Then,an algorithm based on node contribution is proposed.The algorithm is based on the CN and RA algorithms,and introduces the concept of node contribution,and analyzes the influence of common neighbors and their information on the prediction of unknown links.The algorithm makes up for the above two algorithms to consider only the deficiency of single attribute,and further study the number of neighbors and the contribution of node information to the predicted nodes.Finally,the simulation performance of the two algorithms proposed in this paper is verified by simulation experiments,and the comparative analysis of the classicalalgorithms is given.