Link prediction algorithm based on the initial information contribution of nodes

Many link prediction algorithms have originated from the process of information transmis-sion between nodes in recent years. Despite these algorithms can obtain great prediction results, there may be also some limitations. For instance, the size of the initial information amount of nodes is ignored when these kinds of algorithms are constructed. Aiming at this issue, a link prediction algorithm based on the initial information contribution of nodes is proposed in this paper. First of all, the initial information contribution of nodes is quanti-fied by utilizing some topological information around them and an adjustable parameter. In the next, three ways of bidirectional information transmission between nodes are ana-lyzed. After that, the total information amount that received by two nodes through three ways of information transmission is applied to measure the structural similarity between them, to design the link prediction algorithm. At last, the experimental results on sixteen real-world networks demonstrate that the proposed algorithm has great advantages in effectiveness and robustness, compared with ten mainstream benchmark indices. More than that, in order to verify the application performance of the proposed algorithm in the practical scenario, our algorithm is also employed in some social domains, such as the Facebook and crime networks.(c) 2022 Elsevier Inc. All rights reserved.

Automated summary

This paper proposes a link prediction algorithm based on the initial information contribution of nodes. By quantifying the initial information contribution of nodes and analyzing the ways of information transmission between nodes, an effective link prediction algorithm is designed. Experimental results demonstrate its significant advantages in effectiveness and robustness, as well as its good performance in practical applications.