Case-based support for library reference services

Abstract

This paper reports on the results of the first phase of the development of a knowledge management system that provides assistance to reference librarians. This system would employ the parallel use of expert knowledge and case-based retrieval of similar existing question/answer (QA) pairs. CBR is used to store, organize, and represent these QAs in such a way that they could be offered as suggested solutions. The first phase involves testing the possibility of categorizing reference questions in the knowledge base into case-groups and combining attribute matching with text similarity in the similarity measure. Here, CBR is used in a new field which is not domain-specific, and where the level of variability among the potentially high volume of textual cases can be both major and minor. Questions are used in the similarity measure, and the solutions are the paired answers of the retrieved questions. © Springer-Verlag Berlin Heidelberg 2007.