Network embedding is a method for converting nodes in a network into low dimensional vectors, preserving its structure and the similarities among the nodes. Embedding is widely used in many applications, e.g., social network analysis and knowledge discovery. Because of its wide usage, many studies have been proposed, such as DeepWalk, LINE and node2vec. These works are designed for single-layer networks, however, real world networks often possess not just one, but multiple types of connections. Hence it is more appropriate to represent them as multiplex networks, which consist of multiple layers each of which represents one type of relationship. Embedding multiplex networks is difficult because all layer structures have to be taken into consideration. In this paper, we propose MELL, a novel embedding method for multiplex networks, which incorporates an idea of layer vector that captures and characterizes each layer's connectivity. This method exploits the overall structure effectively, and embeds both directed and undirected multiplex networks, whether their layer structures are similar or complementary. We focus on link prediction tasks and test our method and other baseline methods using five data sets from different domains. The results show that our method outperforms all of the baseline methods for all of the data sets.
各種検索
サポート
ヘルプ