Raqamli Transformatsiya va Sun’iy Intellekt
KLASTERLASH ASOSIDA IJTIMOIY TARMOQ YOZISHMALARINING XAVF DARAJASINI TAHLIL QILISH YONDASHUVI (K=4)
1 Jizzax shaxridagi Qozon federal universiteti filiali
2 Toshkent davlat agrar universiteti
Zenodo DOI: 10.5281/zenodo.20320284
Keywords
Abstract
Ushbu maqolada ijtimoiy tarmoqlarda, xususan Telegram platformasida tarqalgan matnli yozishmalarni xavf darajasiga ko‘ra avtomatik guruhlash masalasi ko‘rib chiqiladi. Tadqiqotning asosiy maqsadi K-means klasterlash algoritmining k = 4 parametri asosida ishlash samaradorligini baholash hamda turli hajmdagi datasetlarda klasterlash sifati va barqarorligini aniqlashdan iborat. Tadqiqot jarayonida 1000, 5000, 10000 va 20000 ta yozishmadan iborat to‘rtta turli hajmdagi ma’lumotlar to‘plami shakllantirildi va ular asosida klasterlar sifati, semantik chegaralar hamda ichki uyg‘unlik darajasi tahlil qilindi. Olingan natijalar kichik hajmdagi datasetlarda klasterlash natijalari yetarli darajada barqaror bo‘lmasligini, o‘rtacha hajmdagi ma’lumotlar bazasida esa klasterlar shakllanish jarayonida ekanini ko‘rsatdi. 10000 ta yozishmadan iborat datasetda k = 4 parametri bilan eng toza, aniq va barqaror semantik guruhlanish hosil bo‘lgani aniqlandi. 20000 ta yozishmadan iborat datasetda esa semantik variativlikning ortishi klasterlar barqarorligining pasayishiga olib keldi. Tadqiqot natijalari K-means algoritmi asosida olingan klasterlash natijalari keyingi bosqichda transformer modellarini (XLM-R, mBERT, BERTbek) o‘qitish uchun samarali belgilash bazasini yaratishda muhim ahamiyatga ega ekanini ko‘rsatadi. Ushbu yondashuv ijtimoiy tarmoqlarda xavfli kontentni avtomatik aniqlash va monitoring qilish tizimlarini takomillashtirishda amaliy ahamiyat kasb etadi.
References
- MacQueen J. Some methods for classification and analysis of multivariate observations // Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability. — Berkeley, 1967. — P. 281–297.
- Mikolov T., Chen K., Corrado G., Dean J. Efficient estimation of word representations in vector space // arXiv preprint. — 2013. — arXiv:1301.3781.
- Uzbek Language BERT (BERTbek) Model Description. HuggingFace Model Card. — 2023. — URL: https://huggingface.co (murojaat sanasi: 2025.01.15).
- Conneau A., Khandelwal K., et al. Unsupervised cross-lingual representation learning at scale // Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL). — Seattle, 2020. — P. 8440–8451.
- Devlin J., Chang M.-W., Lee K., Toutanova K. BERT: Pre-training of deep bidirectional transformers for language understanding // Proceedings of NAACL-HLT. — Minneapolis, 2019. — P. 4171–4186.
- Han J., Kamber M., Pei J. Data Mining: Concepts and Techniques. — 3rd ed. — San Francisco: Morgan Kaufmann, 2011. — 703 p.
- Srivastava A., Sahami M. Text Mining: Classification, Clustering, and Applications. — Boca Raton: Chapman & Hall/CRC, 2009. — 344 p.
- Aggarwal C. C., Zhai C. Mining Text Data. — New York: Springer, 2012. — 543 p.
- Raschka S., Mirjalili V. Python Machine Learning: Machine Learning and Deep Learning with Python, scikit-learn, and TensorFlow. — Birmingham: Packt Publishing, 2019. — 622 p.