The article discusses two methods for extracting foreign words from Yakut texts. Foreign words refer to non-integrated lexical units, which have not been adapted to Yakut orthography and are therefore written as in the original language. Based on the fact that most foreign words in Yakut texts come from the Russian language, it is assumed that they have a particular form by which they can be distinguished from the Yakut word forms. The first method reviewed here is based on rules. It implements an algorithm that detects letter combinations that are foreign to the Yakut language. The second method applies a statistical approach to model and differentiate Yakut and Russian letter combinations. The effectiveness of both methods in extracting Russian foreign words is compared with the results of manual highlighting performed by Russian speakers on 6 Yakut texts. This work is a continuation of the article “Identification of Russian borrowings in Yakut texts”, published in “Computer Linguistics and Computational Ontologies. Number 5 (Proceedings of the XXIV Joint International Conference "Internet and Modern Society, IMS-2022.
Original languageEnglish
Pages (from-to)26-34
JournalInternational Journal of Open Information Technologies
Volume10
Issue number11
StatePublished - 2022

ID: 104746137