Sentimental Manipulation on Social Media Data by Bidirectional Encoder Denotation by Transformers (BERT)

View at publisher Summarise with VIRAT
Authors: Kalaiyarasi V, Satish N, Saranya. E, Siva Kumar A, Dhanush P, A. Vigneshwaran
Cited by: 0
References: 20

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The proposed approach analyses social media sentiment using an advanced deep learning model for natural language understanding, Bidirectional Encoder Representations from Transformers (BERT). The system handles unstructured social media text using a data preprocessing pipeline, followed by BERT tokenization and fine-tuning for sentiment classification. The model is evaluated by calculating its precision, recall, accuracy, and$\mathbf{F 1}$-score. The model's capability of understanding latent contextual relationships in texts is integral in detecting the sentiments, jargon, and sarcasm in social media texts. The analysis shows that BERT-based models' sentiment analysis is more accurate with 98.7 % and robust in comparison to traditional machine learning models, which clearly shows its capability for real-time performance of sentiment analysis and opinion mining from various social media platforms.

Fields of Study & Subjects

Sentiment Analysis and Opinion Mining Topic Modeling Hate Speech and Cyberbullying Detection

Publication details

Journal / source
International Conference on Intelligent Sustainable Systems (ICISS)
Pages
1685–1690
Publication date
4 March 2026
Language
EN