Voting for related entities

Entity search is an emerging research topic in Information Retrieval, where the goal is to rank not documents, but entities in response to a given query. A particularly challenging example of this search scenario is when a user's underlying information need is for a list of entities related to a given entity, represented in the query. In this paper, we propose to tackle this problem as a voting process, by considering the occurrence of an entity among the top ranked documents for a given query as a vote for the existence of a relationship between this and the entity in the query. Our proposed approach is evaluated using a large Web test collection, in the context of the TREC 2009 Entity track. The results attest the effectiveness of our approach when compared to the top participants at TREC, with unparalleled gains in terms of recall. Moreover, through a comprehensive failure analysis, we uncover important issues to be considered when tackling this new search scenario and draw valuable insights towards achieving an effective related entity search performance.