This area is primarily concerned with using recommender systems to help students and (junior) researchers locate useful material that they might not otherwise find, or would find much harder to discover.
Projects working in this area
It is recommended that in year 2, JISC fund additional work in the area of recommender systems for resource discovery.
In particular work is needed in the following areas:
- Investigation and implementation of appropriate algorithms. This should look at existing algorithms in use and their broader applicability. We advise that this should include statisticians with experience in areas such as pattern analysis and recommender systems.
- Investigation of the issues and tradeoffs inherent in developing institutional versus shared service recommender systems. For instance, there are likely to be at least some problems associated with recommending resources that are not available locally.
- Investigating and trialling the combination of activity data with rating data. In doing this there need to be acknowledgement that users are very frequently disinclined to provide ratings, and that ways to reduce barriers to participation and increase engagement with rating processes need to be discovered in the context of the system under development and its potential users.
- RISE - try elsewhere using either the software, or if with a different VLE and LMS, then the methods and algorithms.
- SALT - try elsewhere, either discipline based or at other institutions. SALT may also help to enhance the impact of COPAC. A similar approach could also be tried as a shared service, using activity data from a representative sample of university libraries.
- OpenURL - Run a trial to use OpenURL data to enhance existing or build new recommender systems within one or more institutions.