Ontology-based classification of unstructured information

The area of knowledge management (KM) has been addressed with a considerable amount of research in order to develop concepts and technologies for the retrieval of information and knowledge out of a set of heterogeneous data sources. Especially when we deal with files which contain unstructured information, i.e. documents, it is still a huge challenge to classify them automatically into certain domain-dependant categories. Therefore, this paper describes an application and the underlying concepts which are used for a classification based on the available metadata of files, whereas the classification categories can be found in form of ontology classes. This paper discusses experiences and challenges during the implementation with special regard to ontology-based classification algorithms, the underlying framework as well as the importance metadata quality.