Network Together: Node Classification via Cross-Network Deep Network Embedding

Network embedding is a highly effective method to learn low-dimensional node vector representations with original network structures being well preserved. However, existing algorithms are mostly developed for single network, which fail generalized feature across different networks. In this paper, we study cross-network classification problem, aims at leveraging the abundant labeled information from source help classify unlabeled nodes in target network. To succeed such task, transferable features should be learned end, novel deep (CDNE) model proposed incorporate domain adaptation into so as label-discriminative and network-invariant representations. On one hand, CDNE leverages capture proximities between within by mapping more strongly connected have similar latent other attributes labels leveraged networks making same aligned Extensive experiments been conducted, demonstrating that significantly outperforms state-of-the-art classification.