Intelligent information retrieval:: Diagnosing information need.: Part II.: Uncertainty expansion in a prototype of a diagnostic IR tool

This article is part II of a three part series. In part I we described the theoretical framework for developing an "intelligent information retrieval" tool, based on three principles from medical diagnosis theory. In part II, the present article, we outline a prototype of an "intelligent" IR tool, whose purpose is to facilitate information access for an undergraduate seeking information for a history term paper. Our objective is to create a tool that will (i) draw-out the undergraduate's query to the information system by taking the student through the task of doing the term paper and (ii) diagnose the student's information need by measuring his or her degree of topic integration. The degree of integration indicates a class of information need. The classes of information need are based on Kuhlthau's six stage information search process (ISP) model (each stage is a separate information need, demanding different information to satisfy it). The measurement instrument is based on (i) principles from Shannon's mathematical theory of communication and (ii) principles of uncertainty expansion and reduction from differential diagnosis theory. (C) 1998 Elsevier Science Ltd. All rights reserved.