STUN: querying spatio-temporal uncertain (social) networks

In this paper, we consider the problem of social networks whose edges may be characterized with uncertainty, space, and time. We propose a model called spatio-temporal uncertain networks (STUN) to formally define such networks, and then we propose the concept of STUN subgraph matching (or SM) queries. We develop a hierarchical index structure to answer SM queries to STUN databases and show that the index supports answering very complex queries over 1M+ edge networks in under a second. We also introduce the class of STUNRank queries in which we characterize the importance of vertices in STUN databases, taking space, time, and uncertainty into account. We show query-aware and query-unaware versions of STUNRank as well as alternative ways of defining it. We report on an experimental evaluation of STUNRank showing that it performs well on real world networks.