Attention‐based network embedding with higher‐order weights and node attributes

Network embedding aspires to learn a low-dimensional vector of each node in networks, which can apply diverse data mining tasks. In real-life, many networks include rich attributes and temporal information. However, most existing approaches ignore either information or network attributes. A self-attention based architecture using higher-order weights for both static attributed is presented this article. random walk sampling algorithm on capture topological features presented. For the incorporates first-order k-order weights, attribute similarities into one weighted graph preserve networks. previous snapshots containing nodes graph. addition, utilises damping factor ensure that more recent allocate greater weight. Attribute are then incorporated features. Next, authors adopt advanced architecture, Self-Attention Networks, representations. Experimental results classification link prediction reveal our proposed approach competitive against state-of-the-art baseline approaches.