![Scientific and Technical Journal
of Information Technologies, Mechanics and Optics](/images/mag-ntv.png)
Journal
Scientific and Technical Journal of Information Technologies, Mechanics and Optics
UDK519.68
Issue:1 (113)
Download PDF0 Kbyte
The paper considers the problem of Chinese Open Relation Extraction represented in a form of subject-predicate-object. In contrary to well-known multi-phase methods including word segmentation, part-of-speech tagging, and syntactic analysis, we propose a role approach to detection of parts of sentences without preliminary word segmentation. The key idea is to use syntactic words, prepositions and postpositions as part of speech and member of sentence attributes. Coupled with a small dictionary, it is enough for facts extraction by a query. The experiments conducted on a real technical text show satisfactory results comparable to a traditional approach