Graph Learning Dual Graph Convolutional Network For Semi-Supervised Node Classification With Subgraph Sketch

arXiv:2404.12724v1 Announce Type: new
Abstract: In this paper, we propose the G raph Learning D ual G raph Convolutional Neural Network called GLDGCN based on the classical Graph Convolutional Neural Network by introducing dual convolutional layer and graph learning layer. We apply GLDGCN to the semi-supervised node classification task. Compared with the baseline methods, we achieve higher classification accuracy on three citation networks Citeseer, Cora and Pubmed, and we also analyze and discussabout selection of the hyperparameters and network depth. …

abstract apply arxiv classification convolutional neural network cs.lg graph graph learning layer network neural network node paper semi-supervised type