sibsutisВестник СибГУТИThe Herald of the Siberian State University of Telecommunications and Information Science1998-6920СибГУТИ10.55648/1998-6920-2025-19-3-122-138sibsutis-987Research ArticleСтатьиРешение задачи классификации слухов в интернет-новостяхSolving the problem of rumors classification in online newshttps://orcid.org/0009-0004-5786-8107ХудобинАлександр ДмитриевичKhudobinAlexander Dmitrievich

Студент 2 курса магистратуры Воронежского государственного университета, факультет — Прикладной математики, информатики и механики, направление — прикладная информатика

A second-year master's student at Voronezh State University, Faculty of Applied Mathematics, Informatics and Mechanics, majoring in Applied Informatics

2001kad@mail.ruhttps://orcid.org/0000-0003-0040-5764ВоронинаИрина ЕвгеньевнаVoroninaIrina Evgenievna

д-р. техн. наук, доц., профессор кафедры программного обеспечения и администрирования информационных систем Воронежского государственного университета

irina.voronina@gmail.comВоронежский государственный университетРоссияVoronezh State UniversityRussian Federation202530092025193Вестник СибГУТИ122138Copyright © Худобин А.Д., Воронина И.Е., 20252025Худобин А.Д., Воронина И.Е.Khudobin A.D., Voronina I.E.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://vestnik.sibsutis.ru/jour/article/view/987

Рассматривается подход к решению задачи классификации слухов в новостях на основе продукционных правил. Непроверенная информация, появляющаяся на новостных сайтах, имеет характер информационного мусора и способна в отдельных случаях нанести существенный вред потребителям. Решаемая задача носит нетривиальный характер, актуальна и не имеет стандартного решения.

The article considers an approach to solving the problem of classifying rumors in thenews based on production rules. Unverified information appearing on news sites has the natureof information garbage and is capable of causing significant harm to consumers in some cases.The problem being solved is non-trivial, relevant and has no standard solution.

слухиклассификацияLLMмашинное обучениенейронные сетиTF-IDFBag-of-wordsWord2VecN-граммыпродукционные правилаDeepSeek-R1e5-largerumorsclassificationLLMmachine learningneural networksTF-IDFBag-of-wordsWord2VecN-gramsproduction rulesDeepSeek-R1e5-largeReferencesЛингвистическая экспертиза в трудах Воронежской ассоциации экспертов-лингвистов : монография / Ж. В. Грачева [и др.] ; под ред. М. Е. Новичихиной, А. В. Рудаковой. – Воронеж : Издательский дом ВГУ, 2023. – 357 с.Gracheva Zh. V. Lingvisticheskaja jekspertiza v trudah Voronezhskoj associacii jekspertov-lingvistov : monografija [Linguistic expertise in the works of the Voronezh Association of Linguistic Experts: monograph]. Voronezh, Izdatel'skij dom Voronezhskogo gosudarstvennogo universiteta, 2023. 357 p.Худобин А. Д. TF-IDF, Bag-of-words, Word2Vec и N-граммы для решения задачи классификации слухов в новостях / А. Д. Худобин, И. Е. Воронина // Информатика: проблемы, методы, технологии. – 2024. – С. 1285–1294.Khudobin A. D., Voronina I. E. TF-IDF, Bag-of-words, Word2Vec i N-grammy dlja reshenija zadachi klassifikacii sluhov v novostjah [TF-IDF, Bag-of-words, Word2Vec and N-grams for News Rumors Classification Problem]. Informatika: problemy, metody, tehnologii, 2024, pp. 1285–1294.Осетрова Е. В. Слухи в парадигме лингвистической генристики / Е. В. Осетрова // Жанры речи. – 2015. – Т. 12, № 2. – С. 80–89.Osetrova E. V. Rumors in the paradigm of linguistic genristics. International Journal “Speech Genres”, 2015, vol. 12, no. 2, pp. 80–89.Osetrova E. V. Rumours as a Subject of Scientific Analysis: Social Psychology, History, Philology / E. V. Osetrova // Journal of Siberian Federal University. Humanities & Social Sciences. – 2013. – № 9. – С. 1265–1280.Osetrova E. V. Rumours as a Subject of Scientific Analysis: Social Psychology, History, Philology. Journal of Siberian Federal University. Humanities & Social Sciences, 2013, no. 9, pp. 1265–1280.Русакова И. Б. Концепт слух и его репрезентация в русских и английских пословицах / И. Б. Русакова // Международный аспирантский вестник. Русский язык за рубежом. – 2024. – № 1. – С. 60–67.Rusakova I. B. Koncept sluh i ego reprezentacija v russkih i anglijskih poslovicah [The concept of hearing and its representation in Russian and English proverbs]. Mezhdunarodnyj aspirantskij vestnik. Russkij jazyk za rubezhom, 2024, no. 1, pp. 60–67.Иванова С. В. Жанр светских слухов в дискурсе англоязычных массмедиа / С. В. Иванова, Г. Ш Хакимова // Russian Journal of Linguistics. – 2020. – Т. 24, № 2. – С. 386–418.Ivanova S. V, Khakimova G. Celebrity gossip as a genre in English‐language mass media discourse. Russian Journal of Linguistics, 2020, vol. 24, no. 2, pp. 386–418.Khakimova G. Rumour Text Constructing Techniques In Media Discourse: Case Study Of Gossip Columns / The European Proceedings of Social and Behavioural Sciences. – 2019. – С. 346–357.Khakimova G. Rumour Text Constructing Techniques In Media Discourse: Case Study Of Gossip Columns. The European Proceedings of Social and Behavioural Sciences, 2019, pp. 346–357.Хакимова Г. Ш. Концепт «слух» в фокусе лексикографического анализа. Часть 1 / Г. Ш. Хакимова // Вестник ЮУрГУ. Серия «Лингвистика». – 2016. – № 2. – С. 29–36.Khakimova G. CONCEPT “RUMOUR” IN TERMS OF LEXICOGRAPHIC ANALYSIS. PART 1. Bulletin of the South Ural State University series Linguistics, 2016, vol. 13, no. 2, pp, 29–36Хакимова Г. Ш. Концепт «слух» в фокусе лексикографического анализа. Часть 2 / Г. Ш. Хакимова // Вестник ЮУрГУ. Серия «Лингвистика». – 2016. – № 3. – С. 38–46.Khakimova G. CONCEPT “RUMOUR” IN TERMS OF LEXICOGRAPHIC ANALYSIS. PART 2. Bulletin of the South Ural State University series Linguistics, 2016, vol. 13, no. 3, pp, 38–46Li X. Dice Loss for Data-imbalanced NLP Tasks / X. Li [и др.] // Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. – 2020. – С. 465–476.Pastrevich M. K., Voronina I. E. Avtomaticheskaja klassifikacija rechevoj agressii [Automatic classification of speech aggression]. Aktual'nye problemy prikladnoj matematiki, informatiki i mehaniki, 2023, pp. 1359–1365.Duncan J. The Sensitivity of Word Embeddings-based Author Detection Models to Semantic-preserving Adversarial Perturbations / J. Duncan [и др.] // arXiv preprint arXiv:2102.11917. – 2021. – 22 с.Roumeliotis K. I., Tselika N. D., Nasiopoulos D. K. Next-Generation Spam Filtering: Comparative Fine-Tuning of LLMs, NLPs, and CNN Models for Email Spam Classification. Electronics, 2024. vol. 13, no. 11, pp. 2034–2058.Хобсон Л. Обработка естественного языка в действии / Л. Хобсон, Х. Хапке, К. Ховард. – СПб.: Питер, 2020. – 576 с.Li X. et. al. Dice Loss for Data-imbalanced NLP Tasks. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 2020, pp. 465–476.Mohare R. V. ‘Bag of Words’ to ‘Bag of Concepts’: Improving Text Categorization using SVM / R. V. Mohare [и др.] // Nanotechnology Perceptions. – 2024. – № S6. – С. 419–426.Khudobin A. D., Voronina I. E. Podhod k resheniju zadachi klassifikacii sluhov v internet-novostjah [An approach to solving the problem of rumors classification in online news]. Aktual'nye problemy prikladnoj matematiki, informatiki i mehaniki, 2024, pp. 334–340.Suhasini V. A Hybrid TF-IDF and N-Grams Based Feature Extraction Approach for Accurate Detection of Fake News on Twitter Data / V. Suhasini, Dr. N. Vimala // Turkish Journal of Computer and Mathematics Education. – 2021. – vol. 12, no. 06. – С. 5710–5723.Duncan J. et. al. The Sensitivity of Word Embeddings-based Author Detection Models to Semantic-preserving Adversarial Perturbations. 2021, 22 p., available at: https://arxiv.org/pdf/2102.11917 (accessed 13.04.2025).Cahyani D. E. Performance comparison of TF-IDF and Word2Vec models for emotion text classification / D. E. Cahyani, I. Patasik // Bulletin of Electrical Engineering and Informatics. – 2021. – № 5 – С. 2780–2788.Policyna, E. V., Policyn S. A., Porechnyj A. S., Rykunov A. N. Analiz kachestva raboty i rasshirenie vozmozhnostej instrumentov morfologicheskogo analiza tekstov na russkom jazyke [Analysis of the quality of work and expansion of the capabilities of tools for morphological analysis of texts in Russian]. Vestnik Voronezhskogo gosudarstvennogo universiteta. Serija: Sistemnyj analiz i informacionnye tehnologii, 2023, no. 2, pp. 171–180.Abubakar H. D. Sentiment Classification: Review of Text Vectorization Methods: Bag of Words, Tf-Idf, Word2vec and Doc2vec / H. D. Abubakar, M. Umar, M. A. Bakale // Sule Lamido University Journal of Science & Technology. – 2022. – № 1 & 2. – С. 27–33.Hobson L., Hapke H., Hovard K. Obrabotka estestvennogo jazyka v dejstvii [Natural Language Processing in Action]. Saint Petersburg, Piter, 2020, 576 p.Линькова Г. В. Сравнение методов векторных представлений текстов в задачах классификации вакансий / Г. В. Линькова, Д. С. Ботов // Научное сообщество студентов XXI столетия. Технические науки. – 2018. – № 6 – С. 96–113.Mohare R. V. et. al. ‘Bag of Words’ to ‘Bag of Concepts’: Improving Text Categorization using SVM. Nanotechnology Perceptions, no. S6, pp.419–426.Santos R. F. Long Term-short Memory Neural Networks and Word2vec for Self-admitted Technical Debt Detection / R. F. Santos [и др.] // Proceedings of the 22nd International Conference on Enterprise Information Systems. – 2020. – С. 157–165.Suhasini V., Vimala Dr. N. A Hybrid TF-IDF and N-Grams Based Feature Extraction Approach for Accurate Detection of Fake News on Twitter Data. Turkish Journal of Computer and Mathematics Education, 2021, vol. 12, no. 06, pp. 5710–5723.Devlin J. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding / J. Devlin [и др.] // Proceedings of NAACL-HLT. – 2019. С. 4171–4186.Cahyani D. E., Patasik I. Performance comparison of TF-IDF and Word2Vec models for emotion text classification. Bulletin of Electrical Engineering and Informatics, 2021, no. 5, pp. 2780–2788.Wang L. Multilingual E5 Text Embeddings: A Technical Report / L. Wang [и др.] // arXiv pre-print arXiv:2402.05672. – 2024. – 6 с.Abubakar H. D., Umar M., Bakale M. A. Sentiment Classification: Review of Text Vectorization Methods: Bag of Words, Tf-Idf, Word2vec and Doc2vec. Sule Lamido University Journal of Science & Technology, 2022, no. 1 & 2, pp. 27–33.Bi X. DeepSeek LLM: Scaling Open-Source Language Models with Longtermism / X. Bi [и др.] // arXiv preprint arXiv:2401.02954. – 2024. – 48 с.Lin'kova G. V., Botov D. S. Sravnenie metodov vektornyh predstavlenij tekstov v zadachah klassifikacii vakansij [Comparison of Text Vector Representation Methods in Job Classification Problems]. Nauchnoe soobshhestvo studentov XXI stoletija. Tehnicheskie nauki, 2018, no. 6, pp. 96–113.Santos R. F. et. al. Long Term-short Memory Neural Networks and Word2vec for Self-admitted Technical Debt Detection. Proceedings of the 22nd International Conference on Enterprise Information Systems, 2020, pp. 157–165.Santos R. F. et. al. Long Term-short Memory Neural Networks and Word2vec for Self-admitted Technical Debt Detection. Proceedings of the 22nd International Conference on Enterprise Information Systems, 2020, pp. 157–165.Devlin J. et. al. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding Proceedings of NAACL-HLT, 2019, pp. 4171–4186.Devlin J. et. al. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding Proceedings of NAACL-HLT, 2019, pp. 4171–4186.Wang L. et. al. Multilingual E5 Text Embeddings: A Technical Report. 2024, 6 p., available at: https://arxiv.org/pdf/2402.05672 (accessed 13.04.2025).Wang L. et. al. Multilingual E5 Text Embeddings: A Technical Report. 2024, 6 p., available at: https://arxiv.org/pdf/2402.05672 (accessed 13.04.2025).Bi X. et. al. DeepSeek LLM: Scaling Open-Source Language Models with Longtermism. 2024, 48 p., https://arxiv.org/pdf/2401.02954 (accessed 13.04.2025).Bi X. et. al. DeepSeek LLM: Scaling Open-Source Language Models with Longtermism. 2024, 48 p., https://arxiv.org/pdf/2401.02954 (accessed 13.04.2025).

The authors declare that there are no conflicts of interest present.