Registration Number
0223U004521,
(0123U104013)
, R & D reports
Title
Development of methods of processing and annotating textual information
popup.stage_title
Розроблення теоретико-методичного забезпечення та формування практичних рекомендацій щодо попередньої обробки та анотування текстової інформації
Head
Udovenko Serhii H., д.т.н.
Registration Date
14-11-2023
Organization
Simon Kuznets Kharkiv National University of Economics
popup.description1
Improvement of theoretical and methodological support and formation of practical recommendations for preliminary processing and annotation of text information
popup.description2
Considered important aspects of analysis and development of promising methods of text processing and their practical application in publishing houses and printing enterprises.
The results of the development and practical research of the method of neural network classification of Ukrainian-language text news are given. The accuracy of the constructed neural network for full-text news classification is about 84% on the test sample. A complete set of unique words found in the training set was used as a dictionary.
The results of the development and practical research of the method of classifying the similarity of trademarks based on the textual description are presented. The proposed method made it possible to correctly classify the tested trademarks (using the doc2vec neural network)
The report proposes an approach to automatic annotation of texts using the method of connected word forms. The basis of this method is the modified Ginzburg algorithm, which allows you to highlight the key words of the text, as well as related phrases that have a similar meaning and a common theme. The software-implemented method of connected word forms is included in the Booker-M software module, which allows annotating text documents by category. The resulting annotations are a fairly accurate description of the main meaning of the analyzed text.
Product Description
popup.authors
Teslenko Oleh V.
Udovenko Serhii H.
Luhovyi Bohdan V.
popup.nrat_date
2023-11-14
