Evaluating Relation Retrieval for Entities and Experts
Jianhan Zhu, Arjen P. de Vries, Gianluca Demartini, Tereza Iofciu · 2008
Tremendous progress has been made in terms of retrieval models and user evaluation for expert finding. From 2007, INEX provides the XML Entity Ranking track (INEX-XER) as a forum for the study of entity retrieval, a research area closely related to expert finding. Here, instead of being restricted to finding people (a particular type of entity) with a specified expertise (the topic), any type of entity related to a given topic can be the target of the retrieval system. INEX-XER 2008 proposes a novel entity relation search task, which goes beyond entity retrieval by further establishing relations between entities. Based on the connections between expert and entity retrieval, we propose to explore a tentative expert relation search task in this position paper. Our proposal shows how we can bring expert and entity retrieval research together for developing approaches that could potentially be effective for both. We expect this proposal to inspire contributions to expert finding from other research areas than information retrieval, such as semantic web, information extraction, social network analysis, virtual communities, and question answering.