A0819
Title: Link prediction by exploring common neighborhoods
Authors: Tso-Jung Yen - Academia Sinica (Taiwan) [presenting]
Abstract: Social network analysis aims to establish the properties of a network by exploring the link structure of the network. However, due to concerns such as confidentiality and privacy, a social network may not provide full information on its links. As some of the links are missing, it is difficult to establish the network's properties by exploring its link structure. A method for recovering such missing links is proposed. It is paid attention to a situation in which some nodes have fully-observed links. The method relies on exploiting the sub-network of these "anchor" nodes to recover missing links of nodes that have neighbourhoods overlapping with the "anchor" nodes. It uses a graph neural network to extract information from these neighbourhoods and then applies this information to a regression model for missing link recovery. This method is demonstrated on real-world social network data. The results show that this method can achieve better performance than traditional methods that are solely based on node attributes for missing link recovery.
