Teil eines Buches,
Ontologies
The Oxford Handbook of Computational Linguistics, Kapitel 25, Oxford University Press, Oxford, UK, (2005)
DOI: 10.1093/oxfordhb/9780199276349.013.0025
Zusammenfassung
Ontology refers to the storage of information within a domain, to draw common sense inferences. The expressly linguistic nature of this sort of information would translate it into a lexicon. Traditions dealing with knowledge structuring within ontologies, can be positioned depending on their focus on words/concepts, for different purposes. These are, philosophical tradition, cognitive tradition, artificial intelligence tradition, lexical semantics, lexicography, and information science. Ontologically accumulated knowledge bases can be used to inform structural linguistic analysis, as well as partial understanding. However, most current NLP techniques hardly ever perform full language understanding. While NLP generally seems to be shifting towards inferencing systems that exploit common sense knowledge, small-scale information systems can be enhanced by (re)using more general strands of information. Prospects of convergence of different paradigms have also triggered of efforts to standardize ontological contents. In this chapter, we discuss ontologies and their role in natural language processing (NLP). First the notion of ontology is explained from different paradigms and by reference to some common existing ontologies. Next, we discuss how ontologies are commonly used in NLP.
