2007FallQ3Response

3. Relevance has been a topic of research interest in information science for more than three decades. What is relevance? Why is relevance a central concept in information science? What are the conflicting views of relevance and their implications for information retrieval?

Jody: Relevance is a match between the representation of the retrieved information, and the user's need. Part of the difficulty is in defining what is relevant, and what is not, as this is dependent entirely upon the user's perspective. It may be that topics different than the one in the query are indeed relevant to the search. If the user has not clarified his true information need, the retrieved results may be relevant to the request, but not relevant to the user's need. It is also possible that the user retrieves documents which are topically relevant but which either has already been seen or which the user is violently in disagreement with; in which case, is the document relevant? Some of the problems with relevance is that the order in which documents are presented affects the judgments people make, that judgments change as more information comes in, and that different people judge relevance differently. Another problem is that it is very difficult to determine how many documents in a system are actually relevant to the search, as the query is not likely to retrieve all of them (without retrieving a large number of irrelevant results). Relevance is a central concept in information science because it is central to our role as information mediators. We need to be able to vet information on behalf of our users, and to do so effectively means that we must be able to retrieve as much relevant information as possible in response to their needs.