Machine Translation in Website Localization: Assessing its Translation Quality for Language Learning

Abstract

This study seeks to investigate the quality of translation generated by a popular Machine Translation (MT) platform, viz. Google Translate (GT) in translating the content of a Russian flight charterer’s website from English into Indonesian. To analyze the translation quality produced by MT, we compared the original translation of the website texts translated by a human translator and the machine-translated version. The translation quality is limited to the readability level. The texts serving as the data were collected from the menus of a Russian flight chartering service website which was originally translated by an Indonesian translator. An analysis further discovered that the translation produced by MT is divided into two quality categories: less readable and readable. The less readable translation is caused by MT’s inability to adjust to stylistic forms in Indonesian, failure to do adaptation strategy and in some cases, MT tends to follow the same sentence structure of the original text (source text).

Unveiling Semantic Errors Found in Lexical Translations of Tasya Farasya’s Tiktok Account

“…Up to the present time, the accuracy of translation produced by machine translation is still questionable and debatable (Nasution, 2022). Several studies indicate that MT outputs are still far from being equal to human outputs (Läubli, Sennrich, & Volk, 2018;Läubli et al, 2020;Putri & Setiajid 2021;Utami, Jayantini, & Pratiwi, 2021).…”