طراحي يك پيكره‌ موازي انگليسي/فارسي زبان‌شناسي و استخراج خودكار فرهنگ لغت تخصصي از آن

Designing an English/ Persian Parallel Corpus in Linguistics an‎d Automated Extraction of a Specialized Dictionary from It

زبان شناسي

The globalization of knowledge an‎d the everyday increase of the importance of scientific, scholarly communication necessitate the existence of efficient systems to come up with domain-specific parallel corpora an‎d the later automatic term extraction of them. In this research, we collected the Persian an‎d English abstracts of the available papers in linguistic magazines as well as the Persian-English abstracts of the existing linguistic theses on Iran‎doc, using a Web scraper. The collected data then got segmented an‎d aligned in document, sentence, phrase an‎d word level with the aim of designing a Persian-English parallel corpus in linguistic domain. Moreover, we used an appropriate language model to automatically extract linguistic terms from the developed parallel corpus an‎d build a specialized Persian-English dictionary from it. According to the investigations, the segmentation an‎d alignment quality was reported %92 in sentence level an‎d %96 in phrase an‎d word level. In addition, in a comparison between the available English version of the documents an‎d their translated version using the extracted dictionary, the similarity between the bilingual texts in the test data has reached over %94 from %75.

5

155885