The work is dedicated to the study of Russian electronic thesauri and to the development of automated methods for their creation and evaluation. We suggested new quantitative features for assessing the recall of thesauri; the features take into account the recall of the reflection of common vocabulary, as well as the representation recall of definitions and synonymy relationships. An automatic method for assessing the thesauri recall has been developed; it's based on the use of dictionary data and text corpora. Based on the analysis of lexical resources, an automated method for the sets of synonyms (sunsets) creation is proposed. An automated method of detecting hyperion homonym relations, based on the algorithm of their automatic extraction from the explanatory dictionaries for the Russian language, is proposed. The results of the practical use of all developed methods confirm that they can reduce the amount of expert work needed to build lexical resources.
|出版商||Научно-техническое издательство "Горячая линия-Телеком"|
|州||Published - 2018|
- 16.00.00 LINGUISTICS
Киселев, Ю. А., Мухин, М. Ю., Поршнев, С. В., Берг, Б. Г., & Михайлов, Д. В. (2018). АВТОМАТИЗИРОВАННЫЕ МЕТОДЫ ВЫЯВЛЕНИЯ СЕМАНТИЧЕСКИХ ОТНОШЕНИЙ ДЛЯ ЭЛЕКТРОННЫХ ТЕЗАУРУСОВ: монография. Москва: Научно-техническое издательство "Горячая линия-Телеком".