For this AHRC-funded PhD project, I am working closely with OCLC (the Online Computer Library Centre), a not-for-profit computer service and research organization focusing on library services. Membership of OCLC is open to libraries of all types around the world. One of OCLC's key services is WorldCat - essentially an aggregated catalogue of the holdings of all their member libraries. This currently holds more than a billion individual records. This catalogue is accessible to the public via a web front end (www.worldcat.org), allowing users to search for titles and locate the nearest available library held copy. Users can also view and add content such as reviews and tags.
The ultimate goal of my research is to produce a prototype recommender system that can be added to the WorldCat web interface. A recommender system is an information filtering system that presents users with suggested information items from within a collection. The most widely known examples are probably Amazon’s “people who like this also like…” and the MovieLens film recommendation service. Key to our approach is the adoption of a user-centred methodology. Much of the growing body of recommender systems research is based around technical developments – particularly the algorithms that power recommendations. Our approach will instead attempt to first identify more precisely how a recommender system can best align itself with the needs and circumstances of users. We hope to integrate current models of Human Recommender Interaction with more general theories of Information Behaviour through a series of qualitative studies and observation experiments to be run later this year. The second year of the project will be more technical, and involve analysis of WorldCat query logs and other feedback generation methods. OCLC have committed to providing a software engineer to assist with the production of a prototype system based on our findings, which can then be fully evaluated through user testing.
No comments:
Post a Comment