TERMINOLOGY DATABASE AND MACHINE TRANSLATION

Abstract

The purpose of this article is to describe the approach to standardization of scientific and technical terminology. This approach consists in the direct and reverse translation of the Russian term into a foreign, specifically, English, language and the subsequent reverse translation into Russian. In case of coincidence of the original and the result of a double translation of the term, the latter is supposed to be consistent with the internationally accepted terminology. The Levenshtein distance was used to determine the coincidence / misalignment between the original term and the term resulting from direct and reverse translation. The possibility of choosing the best Russian term among the synonymous or converse phrases is shown. The distribution method using PMI (Pointed mutual information) metrics is applied for assessing the semantic proximity of terms. The results of the research show the prospects of using the machine translation system for standardization of scientific and technical terminology, in particular, in the field of psychology. Standardization in this case is understood as consistency with the international, particularly English terminology.