Enhanced Digital Libraries Through Recommendation: Exploring the Use of Citations, Personal Bibliographies, and Metadata to Synthesize Library Services for Individuals

This proposal explores the use of recommender systems technology for guiding

individuals to parts of a large collection as a tool to help library users and librarians alike. The

proposed research is built upon research showing that recommenders can be effective for helping

researchers with the specific task of finding additional citations germane to an existing document.

Applications will be studied in the context of three key sets of research questions about the design

of new library services: (1) Questions about article metadata the valuable information about authors,

citations, venues, and other related information and how metadata can be mined to better guide library

users. (2) Questions about recommender algorithms for a diverse set of information needs, with particular

interest in algorithms that cross metadata types (e.g., finding people from citations or finding citations

from taxonomy data) and on the design of a modular toolkit of algorithms that can be used by librarians

and other power users to synthesize new recommendation types. (3) Questions about the user interfaces

needed to make such tools accessible to and effective for a variety of different users, from students to

experienced librarians. In the end, the research will make novel contributions to the field of digital

libraries and the research areas of recommender systems and human-computer interaction.