Deep attributed network representation learning of complex coupling and interaction

Zhao Li, Xin Wang, Jianxin Li, Qingpeng Zhang

Abstract: Highlights • We propose a structural role proximity enhancement deep autoencoder, which can effectively preserve the highly nonlinear coupling and interactive network topological structure and attribute information. Furthermore, it can preserve more global and local important potential information by capturing high-order proximity and structural role proximity in the network. • We propose a neighbor optimization strategy to modify the Skip-Gram model, which is used to efficiently and seamlessly integrate the network topological structure and attribute information to improve representation learning performance. • We design two structural role proximity enhancement strategies for deep autoencoder model, namely target enhancement strategy and error enhancement strategy. At the same time, we also design two representation learning output strategies, namely connection output strategy and integrated output strategy. These two types of strategies can make the RolEANE framework reasonably expand to four effective model versions, so that it can choose the optimal solution according to different downstream tasks. • We have verified the effectiveness and stability of the RolEANE model framework through extensive experiments on four real datasets. It can be seen from the experimental results that our proposed model outperforms the state-of-the-art network representation learning methods. On the node classification task, the average performance is improved by 4.52% to 10.28% than the optimal baseline method; on the link prediction task, the average performance is 4.63% higher than the optimal baseline method. Abstract Networks that can describe complex systems in nature are increasingly coupled and interacted, and effective modeling on complex coupling and interaction information is an important research direction of artificial intelligence. Representation learning provides us with a paradigm to solve such issues, but the current network representation learning methods are difficult to capture the coupling and interaction information in complex networks. In this paper, we propose a novel deep attributed network representation learning model framework (RolEANE), which can effectively preserve the highly nonlinear coupling and interactive network topological structure and attribute information. We design two different structural role proximity enhancement strategies for the deep autoencoder in the model framework, so that it can efficiently capture network topological structure and attribute information. In addition, the neighbor-modified Skip-Gram model in our model framework can efficiently and seamlessly integrate network topological structure and attribute information, and the selection of an appropriate representation learning output strategy can significantly improve the final performance of the algorithm. The experiments on four real datasets show that our method consistently outperforms the state-of-the-art network representation learning methods. On the node classification task, the average performance is improved by 4.52%–10.28% than the optimal baseline method; on the link prediction task, the average performance is 4.63% higher than the optimal baseline method.