Text-based Knowledge Acquisition for Ontology Engineering
Uwe Quasthoff, Christian G. Wolff · 2002
This paper describes an approach towards ontology engineering that makes use of text technology for extracting relevant semantic relations from document collections. A short description of corpus characteristics and examples of statistical text analysis results show how input for ontology design can be generated automatically. The Topic Map standard is used as an example for standardised representation of ontologies and a toolset for generating raw Topic Maps is described. Finally practical applications of this approach are described and areas of future research in the refinement of ontology generation are described.