Skip to main content Skip to main navigation


Personalized information filtering for mobile applications

Christian Reuschling; Sandra Zilles
In: Klaus P. Jantke (Hrsg.). Knowledge Media Technologies: First International Core-to-core Workshop. Pages 1-11, Diskussionsbeiträge, Vol. 1, Inst. für Medien- und Kommunikationswiss. 2006.


Recent technological development has enhanced research in the field of pervasive computing for mobile applications. In particular, topics like ad-hoc community building and personalized contextual product offers are of relevance for cellular radio providers. In thi s context, we propose a generic approach to personalized information filtering. This concerns first an appropriate representation scheme fo r the application domain, the user, as well as the information to be filtered. Here we propose to model the domain in an ontology with special weight attributes in RDFS, such that personal interests or r esources (e. g., product descriptions) can be represented as RDF instances of this ontology. Using case based reasoning techniques, we second propose an implementation of a similarity measure between such instances. On the one hand, given a special doma in model, this similarity measure allows for filtering a list of resources according to a person’s inte rests in a way immediately suitable for the intended applications; on the other hand, this similarity m easure is defined generally enough to allow for the comparison of RDF instances in general, with different s pecialized similarity measures depending on the intended semantics of similarity. Third, in addition, the problem of how to maintain represent ations of user interests and resource descriptions in a dynamic domain is addressed briefly.