Information Technology Reference
In-Depth Information
data mining to recommend both digital content at the user's fingertips
and physical content, at the user's local library, in one place; how to
move away from the traditional A&I database experience - 'your result
returns zero hits' - by adding value, flexibility and personalization.
Personalization
Following on from highly successful usage-based recommendations by
commercial services such as Amazon.com, ExLibris uses the bX
recommender service, a collaboration of Ex Libris and the Los Alamos
National Laboratory (LANL),
27
to provide recommendations based on
usage data as an add-on service for customers. This move prompted the
University of Huddersfield to start logging data from its SFX menu
requests to see if there was potential in a similar in-house system.
Although the test is at a very early stage, and will never be able to
utilize the sheer amount of data generated by ExLibris, there is enough
data to make recommendations at journal level.
28
The quality of these
recommendations is now being tested.
Low usage or non-usage of resources is a perennial problem in
libraries. However, the use of anonymized personal data may be of use
in encouraging others to use systems; e.g. a first-year psychology
student would be able to log in to the university portal to check e-mail,
university virtual learning environment (VLE) etc.; the portal would
also be able to recommend which books and journal articles to use for
the course, based on previous undergraduates' experience, without the
student ever needing to enter the library or use traditional methods of
resource discovery. This may be seen as dumbing down or spoon-
feeding; however, if it addresses the needs of the non-user by
introducing resource discovery through recommendation or 'accidental
discoveries', then there is benefit in such a system (King, 2008).
Privacy
Each time that your card is swiped, it registers not only the total
amount spent, but exactly what you have spent it on. The supermarkets
can therefore build up a customer profile of each card holder.
(Howell, 1995)
