Recommender Services in Scientific Digital Libraries
In this article we give a survey of the current practice and state-of-the-art of recommender services in scientific digital libraries. With the notable exception of amazon.com and CiteSeer which do not qualify as proper scientific libraries our survey rev
- PDF / 868,162 Bytes
- 41 Pages / 439 x 666 pts Page_size
- 67 Downloads / 194 Views
k and Varian [58] define recommender systems as systems which support the social process of collecting, aggregating, and communicating recommendations for a social group. For universities, recommender systems in scientific digital libraries hold the promise of supporting the time expensive consulting and communication processes of research and education, and thus improve teaching and research processes. Under competitive and economic pressure, the importance of such high-quality feedback processes cannot be overestimated. Unfortunately, our investigations in Sect. 15.2 show that the deployment of recommender systems in scientific digital libraries is still in its infancy. Only a small number of scientific digital libraries is experimenting with extensions of their services by recommender systems. We give a survey of several of these attempts in Sect. 15.2. M. Franke et al.: Recommender Services in Scientific Digital Libraries, Studies in Computational Intelligence (SCI) 120, 377–417 (2008) c Springer-Verlag Berlin Heidelberg 2008 www.springerlink.com
378
M. Franke et al.
A general explanation and analysis of this state of affairs is already offered by the mechanism design problems of recommender systems in [6, 7]. In Sect. 15.3 we transfer the analysis of Avery and Zeckhauser to recommender systems in scientific digital libraries. We link certain types of mechanism design problems with the main classes of recommender services described in [58], and we discuss remedies for the different types of recommender services in the context of scientific digital libraries. In Sect. 15.4, we present and explain the bundle of recommender services which have been developed for the university library of Karlsruhe with the support of the German Research Society (DFG). These recommender services constitute currently the most advanced and comprehensive extension of scientific library services both in scale and in scope of the services. Finally, the recent development in social computing which led to the revival of the internet boom (e.g., flickr, del.icio.us, youTube, and mySpace) adds a new dimension to scientific library recommender services which we discuss in Sect. 15.5.
15.2 A Survey of Recommender Systems in Major Digital Libraries When browsing through the open public access catalogs (OPACs) of Europe’s national libraries which are members of The European Library (formerly Gabriel, the portal of European national libraries that has been funded by the European Union) and the OPAC of the Library of Congress, we could not find a single operational recommender system. However, several scientific digital libraries are already experimenting with such systems. Our report on these systems considers scientific digital libraries, scientific projects, and the most relevant commercial application, amazon.com. For our purposes, the most important classification dimensions for recommender systems which we use in this section are: •
•
Explicit (ratings or reviews) vs. implicit (behavior-based or content-based) recommendations. Explicit recommendation
Data Loading...
