Towards Evaluating the Impact of Ontologies on the Quality of a Digital Library Alerting System

Abstract

Abstract. Advanced personalization techniques are required to cope with novel challenges posed by attribute-rich digital libraries. At the heart of our deeply personalized alerting system is one extensible preference model that serves all purposes. In this paper we focus on ontology and quality assessment in conjunction with our search technology Preference XPath and XML-based semantic annotations of digital library multimedia objects. We evaluate the impacts of automatic query expansion by ontologies by embedding our alerting system PNews as a black box or a glass box in a test lab. It changes configuration parameters on its own, feeds test cases to P-News, compares the results of different configurations, and stores the result set for further evaluations. The most important indications of this work in progress are: The use of ontologies improves the quality of the result set, generates further results of higher quality, and implies the use of knowledge to reduce a loss of focus.