Hybrid Graph Convolutional Networks for Semi-Supervised Classification
Abstract— In recent years, Graph Convolutional Network (GCN) have been successfully applied to many graph classification problems. It has the capability to learn many types data that Convolutional Neural Networks (CNN) cannot handle, such as irregular data. However, we found that GCN can not completely capture the graph structure information and especially for inference on data efficiently. In this paper, we analyze the advantages and disadvantages of several models and propose two different methods of combining models. Based on that, we propose a new model by using ensemble learning Based on GCN. This model has the ability to capture the advantages of multiple models. Finally, we conduct our experiment on several datasets, and the experimental results show that our approach is effective.
Index Terms— Graph Neural Network, Representation Learning, Data Mining.
Dongyang Bao, Wenxin Hu
School of Computer Science and software Engineering, East China Normal University, CHINA
Information Technology services, East China Normal University, CHINA
Cite: Dongyang Bao, Wei Zheng, Wenxin Hu, "Hybrid Graph Convolutional Networks for Semi-Supervised Classification," Proceedings of 2019 the 9th International Workshop on Computer Science and Engineering, pp. 108-116, Hong Kong, 15-17 June, 2019.