This paper presents AL-TERp (Arabic Language Translation Edit Rate - Plus) an extended version of machine translation evaluation metric TER-Plus that supports Arabic language. This metric takes into accounts some valuable linguistic features of Arabic like synonyms and stems, and correlates well with human judgments. Thus, the development of such tool will bring high benefits to the building of machine translation systems from other languages into Arabic that its quality remains under the expectations, specifically for evaluation and optimization tasks.
CITATION STYLE
El Marouani, M., Boudaa, T., & Enneya, N. (2017). Al-terp: Extended metric for machine translation evaluation of Arabic. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10260 LNCS, pp. 156–161). Springer Verlag. https://doi.org/10.1007/978-3-319-59569-6_17
Mendeley helps you to discover research relevant for your work.