Computing and Informatics

Network representation learning has attracted widespread attention as a pre-processing process for some machine learning and deep learning tasks. However, most existing methods only consider influence of nodes' low-order neighbors to represent them. Either nodes' high-order neighbor or the intrinsic characteristic attributes of nodes are ignored, leading to the effect of network representation learning that needs to be improved. This paper proposes a novel model named Structure Enhanced Graph Convolutional Network (SEGCN) to address these limitations. SEGCN consists of the following components, i.e., the network structure enhancement to transform weak relationship into strong relationship, the node feature aggregation to fuse high-order neighbor information. Hence, the SEGCN model can simultaneously integrate network structure information, attribute information, and high-order neighbor relationships together. Experimental results for node classification and node clustering on six datasets show that SEGCN achieves better effectiveness and efficiency than state-of-the-art baselines.