A Technique to Enhance the Process of Handling Gender Agreement Requirement in English-Arabic Machine Translation

Abstract

In this paper we highlight some difficulties in machine translation especially from English to Arabic; and focus on the agreement requirements between subject and verb, as well as between noun and adjective. We proposed and designed an approach to enhance the translation quality in English-Arabic translators; the approach based on the identification of the exact gender of the nouns before the translation process; the target nouns are those whom used for both masculine and feminine in the source language such as driver, programmer, teacher, and so on. We applied and tested the proposed approach on two translators, and the results showed a significant enhancement of the translation output. The proposal is flexible and scalable, it is a rule-based approach, and can be applied on some other features and/or languages.