The Smart Subjects project aims to increase the discovery and use of library resources through the sensible deployment of lightweight resource recommendation services.
Smart Subjects is built on top of large locally developed subject indexes consisting of rich topical keyword content. Since these services use arbitrary text as input to generate recommendations, these services can be deployed flexibly and easily in a variety of contexts. The first available Smart Subjects service is called Related Subjects and recommends a short list of library subject portal pages related to a user's keyword search
How We Did It
The subject index source content currently in use consists of NCSU course catalog descriptions and publication titles extracted from the NCSU Faculty Publications Database. The subject indexes themselves are created using Solr. The subject identification script and current recommendation services are written in Ruby and integrated as a searcher in the QuickSearch application.
Publications and Presentations
Sierra, T. (2007). "Smart Subjects: Application Independent Subject Recommendations",Code4Lib 2007, Athens, Georgia, February 28, 2007.
Sierra, T. (2006). "Indexing institutional data to promote library resource discovery",Proceedings of the 6th ACM/IEEE-CS joint conference on Digital libraries, 362-362.
Sierra, T. (2006). Indexing Institutional Data to Promote Library Resource Discovery", poster for Joint Conference on Digital Libraries 2006, Chapel Hill, North Carolina. [poster]