The article is devoted to the problem of adequacy and equivalence of the texts translated via machine translation systems. The purpose of the study is to analyze existing machine translation technologies, identify the main errors in the translation of the texts of various subjects, and select the translator which does the highest quality translation of various thematic texts. Two main machine translation technologies have been in focus of the research: a rule-based translation technology (Rule-Based Machine Translation, RBMT) and a statistical translation technology (Statistical Machine Translation, SMT). It has been found out that each technology has both advantages and disadvantages. Among all the studied translation systems, namely Translate.ru (PROMT), Trident Software (Pragma), SYSTRANet, Babylon, Google Translate and Yandex.Perevod, Yandex has proved to be the most successful to complete the translation task regardless of the subject of the translation. Furthermore, it translates various lexical units quite well and confidently copes with grammatical constructions. As it has been found out, Google Translate is inferior to Yandex in the translation of lexical units, especially thematic ones, but has almost the same indicators regarding grammatical correctness. In the third place is the PROMT translator, which translates grammatical constructions well, but has problems with translating thematic vocabulary. The conclusion that can be derived from the research is that we have the most reason to advise Yandex.Perevod to use for translating the texts of different subjects. Despite of the fact that a genuine solution to the problem of machine translation has not yet been found, the development of new scientific theories, modern achievements in the field of Computer Science, Programming, and Linguistics give hope that it will be possible to satisfactory solve this task in the immediate future.
展开▼
