Using Machine Translation English - Arabic Procedures and Challenges - A Systematic Review

Abstract

Recent developments in machine translation (MT) have made it easier to move text across languages, particularly those with limited resources, such as Arabic. One of the most frequently spoken languages in the world is Arabic; hence, the scientific community has recently become quite interested in the problem of Arabic Machine Translation (AMT). The purpose of this assessment was to evaluate the methods used for machine translation from English to Arabic, identify problems, and investigate possible fixes. The exploration encompassed six databases: Science Direct, Google Scholar, Emerald, Scopus, IEEE, and Web of Science. Exclusion criteria involved Studies lacking Arabic data collection or those unrelated to qualitative research were excluded. Twenty studies met the requirements, with 14 opting to translate all transcripts into English before the analysis. Conversely, six studies transcribed and analyzed data in Arabic, subsequently translating the results into English or performing a parallel analysis. The rationale behind converting data into English prior to research was to facilitate access to non-Arabic languages and gather their assistance. The findings indicate a prevalent preference among researchers to translate data before examination, demonstrating an awareness of potential meaning loss during translation and its impact on outcomes.