Knowledge Discovery in the Digital Library: access tools for mining science

Issue title: Electronic Tools for Knowledge Discovery and Health

Article type: Research Article

Authors: Dumouchel, Bernard^a | Demaine, Jeffrey^b

Affiliations: [a] Director General, Canada Institute for Scientific and Technical Information, Canada | [b] Researcher, Canada Institute for Scientific and Technical Information, Canada

Abstract: A researcher's interactions with the scientific literature are limited by its overwhelming size, resulting in ever-increasing specialization. Knowledge Discovery is the process of identifying meaningful, unknown relationships between concepts, enabling broader inquiry of the scientific literature. Although largely automated, Knowledge Discovery is an inherently participatory process. While knowledge discovery techniques can uncover hidden relationships in the data, only the user's expertise can give those relationships meaning. As such, these techniques do not replace but rather enhance the scholarly process. The role of moderating the interface between scholarship and the published literature is the core mission of the research library. By enabling researchers to data-mine the scientific literature Knowledge Discovery techniques are a natural extension of the library's role of bringing structure to information and in making that information accessible. CISTI is investigating theories in information science in order to apply knowledge discovery techniques to its collection: Linked Literature Analysis seeks to uncover hidden relationships between concepts that are causally related, and Main Path Analysis identifies the evolution of a research field based on citations. Providing seamless access to e-science makes it possibile to analyze the published literature in a way that augments the scholar's research.

DOI: 10.3233/ISU-2006-26105

Journal: Information Services & Use, vol. 26, no. 1, pp. 39-44, 2006