LEKSIK SINONIMLAR OMBORINI WORD2VEC VA FASTTEXT MODELLARI YORDAMIDA HOSIL QILISH
Khajibaeva S. M.
Annotatsiya
Ushbu maqolada mashinali o‘qitish algoritmlari yordamida o‘zbek tilining leksik sinonimlarni avtomatik ravishda ajratib olish muammosi keltirilgan. Faqat lug‘atga asoslangan leksik sinonimlarni avtomatik ravishda ajratib olish har doim ham to‘liq samarali emas, chunki lug‘atda "sinonimlar" sifatida keltirilgan so‘zlar haqiqiy matnda uchraydigan aniq kontekstga mos kelmasligi mumkin. Agglyutinativ tillar guruhiga kiruvchi o‘zbek tili uchun Word2Vec va FastText usullari asosida olingan natijalar UzWordNet sinsetlari asosida ajratib olindi va taqqoslandi. Tahlil natijalari shuni ko‘rsatdiki, FastText modeli natijalari Word2Vec modeli natijalariga nisbatan sezilarli darajada oshdi. Olingan natijalar shuni ko‘rsatadiki, agglyutinativ tuzilishga ega o‘zbek tili uchun FastText modeli asosida hosil bo‘lgan vektor maydoni sinonim qatorga kiritilgan ko‘plab so‘zlarning semantik yaqinligini muvaffaqiyatli aks ettiradi.
Adabiyotlar
- T. Mikolov, K. Chen, G. Corrado, and J. Dean, “Efficient Estimation of Word Representations in Vector Space,” in Proceedings of the 1st International Conference on Learning Representations (ICLR Workshop), Scottsdale, AZ, USA, 2013.
- P. Bojanowski, E. Grave, A. Joulin, and T. Mikolov, “Enriching Word Vectors with Subword Information,” in Proc. 5th International Conference on Learning Representations (ICLR), Toulon, France, 2017.
- Elov, B. (2024). Matnli ma’lumotlarni Word2Vec, GloVe va FastText chuqur o‘rganish usullari vositasida qayta ishlash. Raqamli Transformatsiya va Sun’iy Intellekt ilmiy jurnali, 2(5), 225.
- K. Madatov and S. Khajibaeva, “Extraction and Data Analysis Basis Words: Case Study on School Corpus,” in 2025 IEEE XVII International Scientific and Technical Conference on Actual Problems of Electronic Instrument Engineering (APEIE), 2025, pp. 1–5.
- N. Abdurahmonova, “UzWordNet: A Lexical-Semantic Database for the Uzbek Language,” in Proceedings of the 10th Global WordNet Conference (GWC), Wroclaw, Poland, 2019, pp. 8–15.