https://doi.org/10.1140/epjb/s10051-023-00495-1
Regular Article - Statistical and Nonlinear Physics
Graph attention network via node similarity for link prediction
1
College of Information Engineering, Yangzhou University, 225127, Yangzhou, China
2
Business School, University of Shanghai for Science and Technology, 200093, Shanghai, China
Received:
31
October
2022
Accepted:
14
February
2023
Published online:
3
March
2023
Link prediction is a classic complex network analytical problem to predict the possible links according to the known network structure information. Considering similar nodes should present closer embedding vectors with network representation learning, in this paper, we propose a Graph ATtention network method based on node Similarity (SiGAT) for link prediction. Specifically, we calculate similar node set for each node in the network by traditional method. The similar nodes and first-order neighbors are assigned an optimal weight through the graph attention network mechanism. Then, we obtain the embedding vectors of nodes with aggregating the information of the similar nodes and first-order neighbor nodes. By incorporating similar nodes, the node embeddings preserve more structure information of the network in low-dimensional embedding space. Finally, the SiGAT represents the links between pairs of nodes with concatenating the node embedding vectors and then trains a classifier to predict novel potential network links. The results of experiments on five real datasets and large-scale artificial datasets, which are the Yeast dataset, Cora dataset, BIO-CE-HT dataset, Human proteins (Vidal) dataset, Human proteins (Stelzl) dataset, and LFR benchmark datasets, show that the SiGAT outperforms the existing popular approaches.
This work was partially supported by the Natural Science Foundation of the Jiangsu Higher Education Institutions of China under Grant No. 22KJD120002.
Copyright comment Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
© The Author(s), under exclusive licence to EDP Sciences, SIF and Springer-Verlag GmbH Germany, part of Springer Nature 2023. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.