<![CDATA[This research investigates the use of Google Neural Matching Translation (GNMT) in translating multilingual signboards in Mataram City into multiple languages. In this paper, the accuracy, readability, and quality of GNMT translations are evaluated to observe their appropriateness in selected well-turned communication environments. Following previous studies, the current study will investigate accuracy, readability, and quality of GNMT translators outputs on translating multilingual signboards in Mataram city compared to its sets produced by human translators. Field work was carried out between June 9 and July 4, 2025 on Jl. Majapahit, Mataram City, West Nusa Tenggara. The population was all multilingual public signboards in Mataram City, while the sample was those along Jl. Majapahit in Mataram. A total of 209 signs (both monolingual and multilingual) were purposively sampled for accuracy, readability and the quality. The results of the study revealed that GNMT achieved a mean score of 4.37 in terms of accuracy and intelligibility as compared to human’s score 5.0. It is good in translating transparent structured texts but does not well with translation cultural idiomatic expressions, where human post editing for adding context make it consistent and informative. On the whole, GNMT delivers 100% OCR success in multilingual signboard translation; it overcomes on clear language but fails to deal with idiomatic expressions and culturally sensitive ones such as name of foods which are indicative that humans are the most suitable part to handle context dependent or culture specific issues.]]>
Copyrights © 2025
