Ontology-Based Semantic Online Classification of Documents: Supporting Users in Searching the Web

In this paper we describe first results of our research on the categorization of texts in order to disambiguate user queries . The dis cussed methods are based on a combination of indexing- and ontology-based information retrieval techniques in an interactive retrieval system. We present an approach to classify search results by mapping them to semantic classes that are defined by the senses of a query term. The criteria defining each class or `sense folder' are derived from the concepts of an assigned ontology, here Multiwordnet. By annotating each element of a result set with the `sense folder' it is classified in, the user gets additional information about each item -- the specific search term is disambiguated with respect to the underlying document -- and can thus decide more easily if the document is relevant or not for his query