aiSelections: Computational Techniques for Matching Faculty Research Profiles to Library Acquisitions

Abstract

As institutional libraries of all sizes adopt more focused acquisitions policies, subject librarians and other selectors will benefit from sophisticated computational approaches that help to identify the monographs, serials, and electronic resources that are likely to receive the most use, thereby reducing interlibrary loan requests, special orders, and unused materials. We describe a pilot study in which data-mining software tools and algorithms were used to summarize faculty biographies, publications, and departmental curricula and to match the resulting profiles to potential monograph selections. We evaluate the effectiveness of these tools by examining the circulation records, interlibrary loan requests, and purchasing receipts from the past several years, noting the computational techniques that are most likely to improve selection accuracy.