A link prediction algorithm based on socialized semi-local information

Link prediction in complex network refers to estimating the possibility of the existence of unknown links between the nodes by selecting the appropriate link prediction algorithms according to the known information of the network topology structure. In this paper, a local link prediction algorithm is presented by considering the interaction between nodes on the basis of traditional resource allocation algorithm and then a link prediction algorithm based on socialized semi-local information is proposed by taking into account the third-order path information on the basis of the above method. Finally, the feasibility and effectiveness of the algorithms are verified by experiment. The experimental results showed both algorithms have high prediction accuracy and achieved a good balance between accuracy and computational complexity. 1553-9105/Copyright © 2014 Binary Information Press.