LTRo: Learning to Route Queries in Clustered P2P IR.

Query Routing is a critical step in P2P Information Retrieval. In this paper, we consider learning to rank approaches for query routing in the clustered P2P IR architecture. Our formulation, LTRo, scores resources based on the number of relevant documents for each training query, and uses that information to build a model that would then rank promising peers for a new query. Our empirical analysis over a variety of P2P IR testbeds illustrate the superiority of our method against the state-of-the-art methods for query routing.