Improved Model Structure based on GCN

With the development of deep learning, important progress has been made in the field of graph neural network. In order to further improve the accuracy of semi-supervised graph node classification based on graph convolutional neural network (GCN), this paper improves the basic GCN by adding a new convolutional layer to increase the depth of the network and improve the accuracy of the model. At the same time, DropEdge mechanism is introduced to improve the generalization ability of the model and enhance the robustness of the model. The model achieves an accuracy of 81.3% on Cora dataset.

關鍵字

Graph Neural Network ； DropEdge ； Cora

參考文獻

Gori M , Monfardini G , Scarselli F . A new model for earning in raph domains[C] //Neural Networks, 2005. IJCNN '05. Proceedings. 2005 IEEE International Joint Conference on. IEEE,2005.

Roweis S.T., Saul L.K. Nonfinear dimensionafity reduction by locally finear embedding[J]. science,2000, 290(5500):2323-2326.

A. Micheli, “Neural network for graphs: A contextual constructive approach,” IEEE Transactions on Neural Networks, vol. 20, no. 3, pp.498–511, 2009

R. Levie, F. Monti, X. Bresson, and M. M. Bronstein, “Cayleynets: Graph convolutional neural networks with complex rational spectral filters,” IEEE Transactions on Signal Processing, vol. 67, no. 1, pp. 97–109, 2017.

Weston, Jason, Ratle, Fred´ eric, Mobahi, Hossein, and Collobert, Ronan. Deep learning via semi-supervised embedding. In Neural Networks: Tricks of the Trade, pp. 639–655. Springer, 2012

國際替代計量

Improved Model Structure based on GCN

全文下載

主題瀏覽