%0 Journal Article
%A 陈勇
%A 翁伟
%A 谢小竹
%T Graph convolutional network combined with random walks and graph attention network for node classification
%D 2024
%R 10.19682/j.cnki.1005-8885.2024.1004
%J 中国邮电高校学报（英文）
%P 1-14
%V 31
%N 3
%X Graph conjoint attention (CAT) network is one of the best graph convolutional networks (GCNs) frameworks,<br>
which uses a weighting mechanism to identify important neighbor nodes. However, this weighting mechanism is<br>
learned based on static information, which means it is susceptible to noisy nodes and edges, resulting in significant<br>
limitations. In this paper, a method is proposed to obtain context dynamically based on random walk, which allows<br>
the context-based weighting mechanism to better avoid noise interference. Furthermore, the proposed context-based<br>
weighting mechanism is combined with the node content-based weighting mechanism of the graph attention (GAT)<br>
network to form a model based on a mixed weighting mechanism. The model is named as the context-based and<br>
content-based graph convolutional network (CCGCN). CCGCN can better discover important neighbors, eliminate<br>
noise edges, and learn node embedding by message passing. Experiments show that CCGCN achieves state-of-the-<br>
art performance on node classification tasks in multiple datasets.<br>
%U https://jcupt.bupt.edu.cn/CN/10.19682/j.cnki.1005-8885.2024.1004