Novelty extraction from special and parallel corpora

• How can corpora assist translators in ways in which resources like translation memories or term databases cannot? Our tests on English, Polish and Swedish parts of the JRC-Acquis Multilingual Parallel show that corpora can provide support for term standardization and variation, and, most importantly, for tracing novel expressions. A corpus tool with an explicit dictionary representation is particularly suitable for the last task. Culler is a tool which allows one to select expressions with words absent from its dictionary. Even if the extracted material may be stained with some noise, it has an undeniable value for translators and lexicographers. The quality of extraction depends in a rather obvious way on the dictionary and text processing but also on the query.

Ämnesord

NATURVETENSKAP  -- Data- och informationsvetenskap -- Datavetenskap (hsv//swe)

NATURAL SCIENCES  -- Computer and Information Sciences -- Computer Sciences (hsv//eng)

Nyckelord

corpus

novelty

terminology

term extraction

translation

dictionary

Computer science

Datavetenskap

Technology

Teknik

Publikations- och innehållstyp

ref (ämneskategori)

kap (ämneskategori)