Сравнительный анализ специальных корпусов текстов для задач безопасности

Вопросы кибербезопасности. 2020. Т. 3. № 37. С. 58–65.

Лаврентьев А. М., Рябова Д. М., Тихомирова Е. А., Fokina A., Chepovskiy A., Sherstinova T.

The purpose of the study: development of a technique for comparing special text corpora for subsequent use in the identification of extremist texts Method: frequency methods and a specificity indicator for text analysis of the corpus platform TXM were used. Results: a methodology for comparative analysis of special text corpora has been developed, which makes it possible to identify implicit links between corpora of heterogeneous texts; the relationships between the vocabulary of illegal and literary texts were revealed; the possibility of using the specificity index to compile a “profile” of a text subcorpus was shown; comparative analysis of the corpus of extremist texts and the corpus of Russian stories of the first third of the twentieth century was made; the relationships between the vocabulary of illegal and literary texts were revealed; the possibilities of using corpus linguistics to study the properties of extremist texts in order to detect illegal Internet resources and messages were shown; the possibilities of using both morphological characteristics of words and pseudo-bases of word occurrences in the analysis of specificity on corpus data have been examined; research results showed that the frequency analysis tools provided by the TXM platform are effective for applications when it is necessary to identify implicit lexical matches between different text corpora.

Keywords: корпусная лингвистика, русская литература, автоматический анализ текста, экстремистские тексты., показатель специфичности