Keywords
Abstract
Ushbu maqolada ijtimoiy tarmoqlardagi matnli hujjatlarni tahlil qilish va tasniflash bo‘yicha mavjud yondashuvlar natijalari solishtirma tahlil qilinadi. Tadqiqotlarda VKontakte, Facebook, Twitter va Instagram platformalarida qo‘llanilgan klasterlash, tasniflash va sentiment tahlil usullari o‘rganildi. K-Means, SVM, Naive Bayes, KNN, LSTM, OLS va ansambl modellari asosida olingan natijalar aniqlik darajasi bo‘yicha taqqoslandi. Shuningdek, TF-IDF, Bag-of-Words, N-gram, POS-tagging va stemming kabi matnni qayta ishlash usullari samaradorligi ko‘rib chiqildi. Tadqiqot natijalari shuni ko‘rsatadiki, katta hajmdagi ma’lumotlar va gibrid modellardan foydalanish aniqlikni sezilarli oshiradi hamda ijtimoiy media monitoringi va xavfli kontentni erta aniqlash tizimlarini yaratishda muhim ahamiyat kasb etadi.
References
Smith J., Brown T. Social Media Text Mining Techniques. – Journal of Data Science, 2018.
Ivanov P. Hybrid Clustering in VKontakte Analysis. – Information Technologies, 2019.
Kumar A., Singh R. Sentiment Analysis using SVM and Naive Bayes. – IEEE Access, 2020.
Boyd D., Ellison N. Social Network Sites: Definition and History. – Journal of Computer-Mediated Communication, 2007.
Chen L. Deep Learning for Social Media Analytics. – Springer, 2021.
VKontakte Data Analysis using SVM and TF-IDF Methods. – Applied Informatics, 2019.
Kohonen T. Self-Organizing Maps. – Springer, 2001.
Facebook User Interaction Patterns Study. – Social Media Research, 2017.
Twitter Data Mining Framework. – ACM Digital Library, 2018.
Instagram Clustering using K-Means. – International Journal of Computer Applications, 2020.
Apache Spark for Social Media Profiling. – Big Data Analytics, 2019.
LSTM-based Sentiment Analysis Model. – Neural Networks Journal, 2020.