ارسل ملاحظاتك

ارسل ملاحظاتك لنا







Homographs in Arabic as a Challenge to Machine Translation: Google Translate as a Sample

المصدر: مجلة آداب الفراهيدي
الناشر: جامعة تكريت - كلية الآداب
المؤلف الرئيسي: Salman, Mohammed Barjes (Author)
المجلد/العدد: مج15, ع52
محكمة: نعم
الدولة: العراق
التاريخ الميلادي: 2023
الشهر: كانون الثاني
الصفحات: 458 - 468
DOI: 10.51990/2228-015-052-050
ISSN: 2074-9554
رقم MD: 1340284
نوع المحتوى: بحوث ومقالات
اللغة: الإنجليزية
قواعد المعلومات: AraBase
مواضيع:
كلمات المؤلف المفتاحية:
Arabic Language | Homograph | Machine Translation | Google Translate | Diacritics | Drop Down
رابط المحتوى:
صورة الغلاف QR قانون
حفظ في:
المستخلص: Machine translation is one of the growing up fields like all the other developing technologies which try to facilitate the matters that are related to human life. In spite of its importance, the translations of this type lack accuracy and need many modifications because they contain various kinds of shortcomings. The current research focuses on one of the problems of machine translation which is the translation of Arabic homographs because the Arabic writing system permits the existence of a number of lexical items to be written with same letters but they have different meanings and different pronunciations according to the diacritics used with letters. The research adopts Google Translate as a service of machine translation as it is the more important nowadays than all other services of this filed. The problem of translating Arabic homographs is represented in two types. The first one is that Arabs tend to type text with no diacritics which makes it difficult to guess which lexical item is used even when they are used in sentences, and the second problem is that Google Translate does not have the ability to distinguish between the homographs even when they are used with diacritics. Three samples are analysed which represent three sets of homographs that are used in sentences to check the compatibility of Google Translate to translate the homographs when they are written with diacritics and without. The results shows that the accuracy is about 30% of the whole translated homographs which is very low rate. The research suggests a solution concerning translating Arabic homographs which is to upload Google Translate with the meanings of all the versions of homographs in relation to diacritics, and when the homograph is used without diacritics, a drop down of words with diacritics should appear to determine which one is intended by the uploader of the text.

ISSN: 2074-9554

عناصر مشابهة