The Necessary But Not Sufficient Role of Ontologies in Applications Requiring Personalized Content Filtering

Contents

We discuss an interactive approach for personalized information retrieval. Applying such approach allows users to create, store, and consume content that assists them in carrying out their current domain-specific tasks. The approach is grounded to the theory of distributed cognition, which emphasizes the involvement of external elements in thinking processes. We extend the scope of distributed cognition theory by applying it in the Semantic Web, where the content consumers and creators can be software agents in addition to humans. This leads us to consider the benefits and limitations of ontologies in the information retrieval process. We formalize and discuss a model for information usefulness determination. The model combines the level of understanding the information with the level of relevance assigned to the information. The understanding is based on whether the agent trying to make use of the information has access to the ontology/ontologies, to which the piece of information in question conforms. Assigning the relevance, instead, is based on the agent’s preferences combined with contextual details currently characterizing the agent. Finally, we discuss a case example, where the information usefulness determination is applied.