Case-based reasoning model applied to information retrieval

This paper concerns work in progress on the application of case based reasoning (CBR) to problems of information retrieval (IR). Many approaches to IR have already been developed. Traditional models involve techniques such as keyword searches and the use of boolean queries which are applied to databases or indexes. Other methods are based on probabilistic or statistical analysis of occurrences of words or phrases, some on structured queries, and yet others work with vector space models or semantic models. Knowledge-based approaches are not uncommon. However, there is little experimental evidence to demonstrate the effectiveness of knowledge-based techniques and they remain a significant challenge for research. Adaptation and learning have long been viewed as crucial parts of an IR system, and this alone suggests that CBR might prove to be a valuable approach to this area. In this paper, a CBR model is proposed as an alternative approach for IR. Close attention is paid to what exactly should constitute a case; and a system is described which starts from a problem description, i.e. a need for information, and a context, which involves both intellectual and institutional features.