Abstract
The article describes an experimental research the aim of which was to study the impact of using three neural machine translation engines (Google Translate, Microsoft Translator, online-translator.eu) on translation quality of an economic text in terms of the number of errors of different types. In order to achieve the aim we solved a number of tasks: formulated a research hypothesis, selected an economic text and three neural machine translation engines, determined the optimal procedure for evaluating the results of translations performed by the selected neural machine translation engines, translated the text by means of the selected neural machine translation engines, assessed the translations according to the established evaluation procedure, analyzed and interpreted the results obtained during the empirical research. The evaluation procedure used in our study was developed by Leonid Chernovaty who suggested establishing penal points for certain types of errors. Thus, he distinguishes three types of errors: first type errors – serious errors that distort the content of the original text or simply omit certain semantic fragments (1,0 penal point is given); second type errors – errors that could theoretically affect the understanding of the content of the original text, that is when the information is transmitted ambiguously, and the translation provides for the possibility of broad interpretation, which may be erroneous (0,5 penal point is given); third type errors – minor errors that do not affect the content of the original text, but which can spoil the impression of the translated text: incorrect punctuation, incorrect spelling, grammatical errors, etc. (0,1 penal point is given). The subject of the research, its purpose and tasks led to the use of such empirical methods as a control translation of an economic text by means of three neural machine translation engines (Google Translate, Microsoft Translator, online-translator.eu), quantitative method of processing experimental data. The obtained results confirmed the hypothesis formulated in the beginning of the study: different neural machine translation engines provide translation of the same economic text of different quality, which can be traced by counting the number of errors in each performed translation. The gap in the results demonstrated by our engines is not crucial, but the translation of the best quality was performed by Google Translate.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.