Developing linguistic ontologies in broad domains

The paper describes a model of a linguistic ontology for automatic document processing in a broad domain, that is a domain comprising thousands of entity classes and unrestricted number of possible relations between them. We present a novel set of relations between concepts, which was specially developed for automated document processing and information retrieval. We propose to use a small set of relations comparable with relations in traditional information-retrieval thesauri. However, we use stricter, ontologically motivated definitions for establishing relations. The utilized relations describe the most sufficient relationships between entities, can be applied for description of various domains. The set of concepts relations is illustrated with the example from newly published linguistic ontology - thesaurus of Russian language RuThes.