MACHINE LEARNING IN ADAPTATION OF EDUCATIONAL TEXTS: LEXICAL ASPECT

This paper presents the results of approbation of machine learning methods for adapting educational texts for the purpose of teaching non-native learners in Russian. The testing was carried out on the text of the textbook on the history of Russia for the 11th grade, in which synonymic substitutions of a number of lexemes were made for words included in the lexical minimum of the A1 level. The preparatory stage of adaptation is described and the classification of errors made in the text as a result of adaptation using machine learning methods is presented.