Fast and scalable algorithms for semi-supervised link prediction on static and dynamic graphs
ECML PKDD'10 Proceedings of the 2010 European conference on Machine learning and knowledge discovery in databases: Part III
Hi-index | 0.00 |
Large bipartite graphs that evolve and grow over time (e.g. new links arrive, old links die out, or link weights change) arise in many settings, such as social networks, co-citations, market-basket analysis, and collaborative filtering. How do we monitor (i) the centrality of an individual node (e.g. which is the most important conference), or (ii) the proximity of two nodes or sets of nodes (e.g. who are the most influential authors with respect to a particular conference)? How can we do this efficiently and incrementally? How can we provide ‘any-time’ answers to interesting queries, with respect to node centrality or proximity? In this paper we propose pTrack and cTrack, which are based on RWR, together with some important modifications to adapt these measures to a dynamic, evolving setting. Additionally, we develop techniques for fast, incremental updates of these measures that allow us to track them continuously as link updates arrive. In addition, we discuss variants of our method that can handle batch updates, as well as place more emphasis on recent links. We demonstrate the effectiveness and efficiency of our methods on several real datasets. Copyright © 2008 Wiley Periodicals, Inc. Statistical Analysis and Data Mining 1: 000-000, 2008