Continuous top-k queries over real-time web streams. (Evaluation de requêtes top-k continues à large-échelle)

The Web has become a large-scale real-time information system forcing us to revise both how to effectively assess relevance of information for a user and how to efficiently implement information retrieval and dissemination functionality. To increase information relevance, Real-time Web applications such as Twitter and Facebook, extend content and social-graph relevance scores with “real-time” user generated events (e.g. re-tweets, replies, likes). To accommodate high arrival rates of information items and user events we explore a publish/subscribe paradigm in which we index queries and update on the fly their results each time a new item and relevant events arrive. In this setting, we need to process continuous top-k text queries combining both static and dynamic scores. To the best of our knowledge, this is the first work addressing how non-predictable, dynamic scores can be handled in a continuous top-k query setting.