Italian-Arabic domain terminology extraction from parallel corpora
Fathi Fawi, Rodolfo Delmonte · Accademia University Press eBooks · 2015
In this paper we present our approach to extract multi-word terms (MWTs) from an Italian-Arabic parallel corpus of legal texts. Our approach is a hybrid model which combines linguistic and statistical knowledge. The linguistic approach includes Part Of Speech (POS) tagging of the corpus texts in the two languages in order to formulate syntactic patterns to identify candidate terms. After that, the candidate terms will be ranked by statistical association measures which here represent the statistical knowledge. After the creation of two MWTs lists, one for each language, the parallel corpus will be used to validate and identify translation equivalents.