kaz44Вестник Университета Шакарима. Серия технические наукиBulletin of Shakarim University. Technical Sciences2788-79953006-0524«Шәкәрім университеті» КеАҚ10.53360/2788-7995-2025-4(20)-25kaz44-2232Research ArticleАВТОМАТИЗАЦИЯ И ИНФОРМАЦИОННЫЕ ТЕХНОЛОГИИ (ОРИГИНАЛЬНАЯ СТАТЬЯ)AUTOMATION AND INFORMATION TECHNOLOGY (ORIGINAL ARTICLE)ОБНАРУЖЕНИЕ ЯЗЫКА НЕНАВИСТИ НА КАЗАХСКОМ ЯЗЫКЕ НА ОСНОВЕ ГЛУБОКОГО ОБУЧЕНИЯ: ГИБРИДНАЯ СИСТЕМА ДЛЯ НАДЕЖНОГО АНАЛИЗА ТЕКСТОВDEEP LEARNING-BASED HATE SPEECH DETECTION IN KAZAKH: A HYBRID FRAMEWORK FOR ROBUST TEXT ANALYSIShttps://orcid.org/0000-0002-1611-1923СултанД.SultanD.

Данияр Султан – PhD, доцент факультета цифровых технологий

050035, Республика Казахстан, г. Алматы, ул. Жандосова, 55 

Daniyar Sultan – PhD, Ac. Associate Professor of the School of Digital Technologies

050035, Republic of Kazakhstan, Almaty, 55 Zhandossova street

daniyar.sultan@narxoz.kzhttps://orcid.org/0000-0002-5508-389XАбдрахмановР.AbdrakhmanovR.

Рустам Абдрахманов – кандидат технических наук, доцент

161205, Республика Казахстан, г.Туркестан, ул. Рабиги Султан Бегим, 14 А

Rustam Abdrakhmanov – Candidate of Technical Sciences, Associate Professor of the Humanities school

161205, Turkistan, Kazakhstan 14 A Rabiga Sultan Begim str.

abdrakhmanov.rustam@iuth.edu.kzhttps://orcid.org/0000-0002-1561-8255АдалыЭшрефAdaliEsref

Эшреф Адалы – PhD, профессор факультета вычислительной техники и информатики

34010, Стамбул

Tursinbay Turymbetov – Candidate of Technical Sciences, Associate Professor of the Humanitiesschool, International University of Tourism and Hospitality

34010, Istanbul

adali@itu.edu.trhttps://orcid.org/0000-0003-0178-8701TурымбетовТ.TurymbetovT.

Турсынбай Турымбетов – кандидат технических наук, доцент

161205, Республика Казахстан, г. Туркестан, ул. Рабиги Султан Бегим, 14 А

Tursinbay Turymbetov – Candidate of Technical Sciences, Associate Professor of the Humanities school

161205, Turkistan, Kazakhstan 14 A Rabiga Sultan Begim str. 

t.turimbetov@iuth.edu.kzhttps://orcid.org/0000-0002-1635-4693БекешоваГ.BekeshovaG.

Гульвира Бауыржановна Бекешова – магистр технических наук, старший преподаватель кафедры Информационной безопасности факультета  информационных технологий 

010000, Республика Казахстан, г. Астана, ул. К. Сатпаева, 2

Gulvira Bekeshova – Senior Lecturer at the Department of Information Security, IT Faculty  

010000, Republic of Kazakhstan, Astana, K. Satpayev Street, 2 

gulvirabauyrzhanovna@gmail.comУниверситет НархозКазахстанNarxoz UniversityKazakhstanМеждународный университет туризма и гостеприимстваКазахстанInternational University of Tourism and HospitalityKazakhstanСтамбульский технический университетТурцияIstanbul Technical UniversityTurkeyЕвразийский национальный университет им. Л.Н. ГумилеваКазахстанL.N. Gumilyov Eurasian national universityKazakhstan20252501202614(20)210219Copyright © Султан Д., Абдрахманов Р., Адалы Э., Tурымбетов Т., Бекешова Г., 20262026Султан Д., Абдрахманов Р., Адалы Э., Tурымбетов Т., Бекешова Г.Sultan D., Abdrakhmanov R., Adali E., Turymbetov T., Bekeshova G.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://tech.vestnik.shakarim.kz/jour/article/view/2232

В данном исследовании представлена новая интеллектуальная система, основанная на методах глубокого обучения, предназначенная для автоматического обнаружения языка ненависти на казахском языке. Особое внимание уделено специфике казахского языка как малоресурсного, где ограниченность лингвистических данных создает значительные трудности при построении надежных моделей. На основе различных онлайн-источников – социальных сетей, форумов и новостных порталов – был сформирован и предобработан многоязычный корпус данных, охватывающий широкий спектр речевых контекстов. Для повышения эффективности предложена гибридная архитектура, включающая свёрточные нейронные сети (CNN), двунаправленные долгосрочные краткосрочные памяти (BiLSTM) и механизмы внимания Transformer. Проведённая оценка по метрикам точности, полноты, F1-критерия и достоверности показала превосходство предложенной модели над традиционными алгоритмами машинного обучения. Результаты исследования вносят значимый вклад в развитие систем автоматической модерации контента, технологий анализа текстов на казахском языке и способствуют формированию более безопасного, инклюзивного и устойчивого цифрового пространства для казахскоязычных пользователей.

This study presents a new intelligent system based on deep learning methods for the automatic detection of hate speech in the Kazakh language. Particular attention is paid to the specific nature of Kazakh as a resource-poor language, where the limited linguistic data poses significant challenges in building reliable models. A multilingual data corpus covering a wide range of speech contexts was generated and preprocessed using various online sources-social media, forums, and news portals. To improve efficiency and generalization performance, a hybrid architecture was proposed, including convolutional neural networks (CNNs), bidirectional long short-term memories (BiLSTMs), and Transformer attention mechanisms. An evaluation using precision, recall, F1-criterion, and accuracy metrics demonstrated the superiority of the proposed model over traditional machine learning algorithms. The results of the study make a significant contribution to the development of automatic content moderation systems and promote the creation of a safer, inclusive, and linguistically sensitive digital space for Kazakh-speaking users.

глубокое обучениеязык враждыказахский языкклассификация текстаобработка естественного языкаdeep learninghate speechKazakh languagetext classificationnatural language processingThis work was supported by the research project «Automatic detection of cyberbullying among young people in social networks using artificial intelligence» funded by the Ministry of Science and Higher Education of the Republic of Kazakhstan. Grant No. IRN AP23488900.ReferencesData-Driven Morphological Analysis and Disambiguation for Kazakh / O. Makhambetov et al // Computational Linguistics and Intelligent Text Processing. – 2015. – Р. 151-163. https://doi.org/10.1007/978-3-319-18111-0_12.Data-Driven Morphological Analysis and Disambiguation for Kazakh / O. Makhambetov et al // Computational Linguistics and Intelligent Text Processing. – 2015. – Р. 151-163. https://doi.org/10.1007/978-3-319-18111-0_12.Assembling the Kazakh Language Corpus / O. Makhambetov et al // in Proc. 2013 Conf. Empirical Methods in Natural Language Processing (EMNLP), Seattle, WA, USA. – 2013. – Р. 1022-1031. [Online]. Available: https://aclanthology.org/D13-1104.Assembling the Kazakh Language Corpus / O. Makhambetov et al // in Proc. 2013 Conf. Empirical Methods in Natural Language Processing (EMNLP), Seattle, WA, USA. – 2013. – Р. 1022-1031. [Online]. Available: https://aclanthology.org/D13-1104.Yessenbayev Z. KazNLP: A Pipeline for Automated Processing of Texts Written in Kazakh Language / Z. Yessenbayev, Z. Kozhirbayev, A. Makazhanov // in Speech and Computer. Switzerland: Springer. – 2020. – Р. 657-666. https://doi.org/10.1007/978-3-030-60276-5_63.Yessenbayev Z. KazNLP: A Pipeline for Automated Processing of Texts Written in Kazakh Language / Z. Yessenbayev, Z. Kozhirbayev, A. Makazhanov // in Speech and Computer. Switzerland: Springer. – 2020. – Р. 657-666. https://doi.org/10.1007/978-3-030-60276-5_63.Document and Word-level Language Identification for Noisy User Generated Text / Z. Kozhirbayev, Z. Yessenbayev, A. Makazhanov // in Proc. 12th Int. Conf. Application of Information and Communication Technologies (AICT), Almaty, Kazakhstan. – 2018. – Р. 1-4. https://doi.org/10.1109/ICAICT.2018.8747138.Document and Word-level Language Identification for Noisy User Generated Text / Z. Kozhirbayev, Z. Yessenbayev, A. Makazhanov // in Proc. 12th Int. Conf. Application of Information and Communication Technologies (AICT), Almaty, Kazakhstan. – 2018. – Р. 1-4. https://doi.org/10.1109/ICAICT.2018.8747138.Yessenbayev Z. KazNLP: A Pipeline for Automated Processing of Texts Written in Kazakh Language / Z. Yessenbayev, Z. Kozhirbayev, A. Makazhanov // in Speech and Computer. LNCS. – 2020. – vol. 12335. – Р. 657-666. https://doi.org/10.1007/978-3-030-60276-5_63.Yessenbayev Z. KazNLP: A Pipeline for Automated Processing of Texts Written in Kazakh Language / Z. Yessenbayev, Z. Kozhirbayev, A. Makazhanov // in Speech and Computer. LNCS. – 2020. – vol. 12335. – Р. 657-666. https://doi.org/10.1007/978-3-030-60276-5_63.Data-Driven Morphological Analysis and Disambiguation for Kazakh / O. Makhambetov et al // in CICLing. – 2015. – vol. 9041. – Р. 151-163. https://doi.org/10.1007/978-3-319-18111-0_12.Data-Driven Morphological Analysis and Disambiguation for Kazakh / O. Makhambetov et al // in CICLing. – 2015. – vol. 9041. – Р. 151-163. https://doi.org/10.1007/978-3-319-18111-0_12.Yessenbayev Z. KazNLP: A Pipeline for Automated Processing of Texts Written in Kazakh Language / Z. Yessenbayev, Z. Kozhirbayev, A. Makazhanov // in SPECOM. – 2020. – vol. 12335. – Р. 657-666. https://doi.org/10.1007/978-3-030-60276-5_63.Yessenbayev Z. KazNLP: A Pipeline for Automated Processing of Texts Written in Kazakh Language / Z. Yessenbayev, Z. Kozhirbayev, A. Makazhanov // in SPECOM. – 2020. – vol. 12335. – Р. 657-666. https://doi.org/10.1007/978-3-030-60276-5_63.Development of CRF and CTC Based End-To-End Kazakh Speech Recognition System / D. Oralbekova et al // in Intelligent Information and Database Systems. – 2022. – vol. 13757. – Р. 519-531. https://doi.org/10.1007/978-3-031-21743-2_41.Development of CRF and CTC Based End-To-End Kazakh Speech Recognition System / D. Oralbekova et al // in Intelligent Information and Database Systems. – 2022. – vol. 13757. – Р. 519-531. https://doi.org/10.1007/978-3-031-21743-2_41.A Comparative Analysis of LSTM and BERT Models for Named Entity Recognition in Kazakh Language: A Multi-classification Approach / D. Oralbekova et al // in Modeling and Simulation of Social-Behavioral Phenomena in Creative Societies (MSBC 2024), CCIS. – 2024. – vol. 2211. – Р. 116-128. https://doi.org/10.1007/978-3-031-72260-8_10.A Comparative Analysis of LSTM and BERT Models for Named Entity Recognition in Kazakh Language: A Multi-classification Approach / D. Oralbekova et al // in Modeling and Simulation of Social-Behavioral Phenomena in Creative Societies (MSBC 2024), CCIS. – 2024. – vol. 2211. – Р. 116-128. https://doi.org/10.1007/978-3-031-72260-8_10.Neurocomputer System of Semantic Analysis of the Text in the Kazakh Language / A. Akanova et al // ACM Trans. Asian and Low-Resource Language Information Processing. – 2024. – vol. 23, № 4. https://doi.org/10.1145/3652159.Neurocomputer System of Semantic Analysis of the Text in the Kazakh Language / A. Akanova et al // ACM Trans. Asian and Low-Resource Language Information Processing. – 2024. – vol. 23, № 4. https://doi.org/10.1145/3652159.Automatic Recognition of Kazakh Speech Using Deep Neural Networks / O. Mamyrbayev et al // in Asian Conf. Intelligent Information and Database Systems. – 2019. – vol. 11432. – Р. 465-474. https://doi.org/10.1007/978-3-030-14802-7_40.Automatic Recognition of Kazakh Speech Using Deep Neural Networks / O. Mamyrbayev et al // in Asian Conf. Intelligent Information and Database Systems. – 2019. – vol. 11432. – Р. 465-474. https://doi.org/10.1007/978-3-030-14802-7_40.End-to-End Speech Recognition in Agglutinative Languages / O. Mamyrbayev et al // in Intelligent Information and Database Systems. – 2020. – vol. 12034. – Р. 391-401. https://doi.org/10.1007/978-3-030-42058-1_33.End-to-End Speech Recognition in Agglutinative Languages / O. Mamyrbayev et al // in Intelligent Information and Database Systems. – 2020. – vol. 12034. – Р. 391-401. https://doi.org/10.1007/978-3-030-42058-1_33.A Comparative Analysis of LSTM and BERT Models for Named Entity Recognition in Kazakh Language: A Multi-classification Approach / D. Oralbekova et al // in MSBC. – 2024. – vol. 2211. – Р. 116-128. https://doi.org/10.1007/978-3-031-72260-8_10.A Comparative Analysis of LSTM and BERT Models for Named Entity Recognition in Kazakh Language: A Multi-classification Approach / D. Oralbekova et al // in MSBC. – 2024. – vol. 2211. – Р. 116-128. https://doi.org/10.1007/978-3-031-72260-8_10.Advanced Implementation of a Multilevel Model for Text Summarization in Kazakh Using Pretrained Models / D. Oralbekova et al // Engineering, Technology & Applied Science Research. – 2025. – vol. 15, № 5. – Р. 26711-26721. https://doi.org/10.48084/etasr.12799.Advanced Implementation of a Multilevel Model for Text Summarization in Kazakh Using Pretrained Models / D. Oralbekova et al // Engineering, Technology & Applied Science Research. – 2025. – vol. 15, № 5. – Р. 26711-26721. https://doi.org/10.48084/etasr.12799.A Comparative Analysis of LSTM and BERT Models for Named Entity Recognition in Kazakh Language: A Multi-classification Approach / D. Oralbekova et al // in MSBC 2024, CCIS. – 2024. – vol. 2211. – Р. 116-128. https://doi.org/10.1007/978-3-031-72260-8_10.A Comparative Analysis of LSTM and BERT Models for Named Entity Recognition in Kazakh Language: A Multi-classification Approach / D. Oralbekova et al // in MSBC 2024, CCIS. – 2024. – vol. 2211. – Р. 116-128. https://doi.org/10.1007/978-3-031-72260-8_10.Development of CRF and CTC Based End-To-End Kazakh Speech Recognition System / D. Oralbekova et al // in ACIIDS. – 2022. – vol. 13757. – Р. 519-531. https://doi.org/10.1007/978-3-031-21743-2_41.Development of CRF and CTC Based End-To-End Kazakh Speech Recognition System / D. Oralbekova et al // in ACIIDS. – 2022. – vol. 13757. – Р. 519-531. https://doi.org/10.1007/978-3-031-21743-2_41.

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