Key points are not available for this paper at this time.
Machine Translation (MT) systems mechanically represent a source language into a destination language while maintaining the originality of context using various Natural Language Processing approaches. The need to assess the translation quality produced by machine translation systems has grown as a result of their increased use. However, due to their complicated morphology and syntax, other low-resource languages might not always be appropriate for or applicable to the existing evaluation measures intended for English and other languages. The technique of evaluating machine translation is known as machine translation evaluation (MTE). In MTE, the amount of similarity and accuracy is determined by contrasting the output produced by the machine translation with the reference translation. The study assesses the metrics on several translation systems using datasets for low-resource languages. The paper advances the field of machine translation (MT) by illuminating appropriate evaluation criteria for low-resource languages. For this research. data from the Scopus database were gathered and analyzed using the keywords "machine translation" AND "evaluation metrics" from 1993 to 2023. Using Scopus databases, this article computes numerous application areas, documents published yearly, and keyword occurrences analysis. This work provides analysis and support for the significance of evaluation measures for different low-resource languages.
Kaur et al. (Mon,) studied this question.