a hybrid deep network representation model for detecting researchers’ communities

Recently, network representation has attracted many research works mostly concentrating on representing the nodes in a dense low-dimensional vector. there exist however, some network embedding methods focusing only on the node structure, and some others considering the content information within the nodes. in this paper, we propose a hybrid deep network representation (hdnr) model that uses a triplet deep network architecture considering both the node structure and the content information for network representation. in addition, the author's writing style is also considered as a significant feature in the node content information. due to the successful application of deep learning in natural language processing (nlp), our model makes use of a deep random walk method in order to exploit the inter-node structures and two deep sequence prediction methods to extract the nodes' content information. the embedding vectors generated in this manner are shown to have the ability of boosting each other for learning the optimal node representation, detecting more informative features, and ultimately a better community detection. the experimental results obtained confirm the efficiency of this model for network representation compared to the other baseline methods.