GRE: hybrid recommendations for NSDL collections

ABSTRACT

Recommendation systems have been proven to reduce the time and effort required by users to find relevant items, but there are only sporadic reports on their application in digital libraries. The General Recommendation Engine (GRE) is composed of the text search system Lucene augmented by the well-understood content based and collaborative filtering techniques and the first application of knowledge based recommendation in digital libraries to recommend items from 22 National Science Digital Library collections. In this study comprised of 60 subjects, the GRE outperformed the baseline system Lucene in all areas of evaluation.

INDEX TERMS

Primary Classification:
  H. Information Systems
  H.5 INFORMATION INTERFACES AND PRESENTATION (I.7)
      H.5.2 User Interfaces (D.2.2, H.1.2, I.3.6)

Additional Classification:
  H. Information Systems
  H.3 INFORMATION STORAGE AND RETRIEVAL
      H.3.7 Digital Libraries

General Terms:
Algorithms, Design, Experimentation, Human Factors, Performance

Keywords:
collaborative filtering, content based, digital libraries, knowledge based recommendation, recommendation systems, text search engine, user interface