A BERT–LSTM–Attention Framework for Robust Multi-Class Sentiment Analysis on Twitter Data

Сохранить в:
Библиографические подробности
Опубликовано в::Systems vol. 13, no. 11 (2025), p. 964-983
Главный автор: Zhang, Xinyu
Другие авторы: Liu, Yang, Zhang, Tianhui, Hou Lingmin, Liu, Xianchen, Guo Zhen, Mulati Aliya
Опубликовано:
MDPI AG
Предметы:
Online-ссылка:Citation/Abstract
Full Text + Graphics
Full Text - PDF
Метки: Добавить метку
Нет меток, Требуется 1-ая метка записи!

MARC

LEADER 00000nab a2200000uu 4500
001 3275564978
003 UK-CbPIL
022 |a 2079-8954 
024 7 |a 10.3390/systems13110964  |2 doi 
035 |a 3275564978 
045 2 |b d20250101  |b d20251231 
084 |a 231636  |2 nlm 
100 1 |a Zhang, Xinyu  |u Department of Computer Science, Rochester Institute of Technology, Rochester, NY 14623, USA; lh1026@rit.edu 
245 1 |a A BERT–LSTM–Attention Framework for Robust Multi-Class Sentiment Analysis on Twitter Data 
260 |b MDPI AG  |c 2025 
513 |a Journal Article 
520 3 |a This paper proposes a hybrid deep learning model for robust and interpretable sentiment classification of Twitter data. The model integrates Bidirectional Encoder Representations from Transformers (BERT)-based contextual embeddings, a Bidirectional Long Short-Term Memory (BiLSTM) network, and a custom attention mechanism to classify tweets into four sentiment categories: Positive, Negative, Neutral, and Irrelevant. Addressing the challenges of noisy and multilingual social media content, the model incorporates a comprehensive preprocessing pipeline and data augmentation strategies including back-translation and synonym replacement. An ablation study demonstrates that combining BERT with BiLSTM improves the model’s sensitivity to sequence dependencies, while the attention mechanism enhances both classification accuracy and interpretability. Empirical results show that the proposed model outperforms BERT-only and BERT+BiLSTM baselines, achieving F1-scores (F1) above 0.94 across all sentiment classes. Attention weight visualizations further reveal the model’s ability to focus on sentiment-bearing tokens, providing transparency in decision-making. The proposed framework is well-suited for deployment in real-time sentiment monitoring systems and offers a scalable solution for multilingual and multi-class sentiment analysis in dynamic social media environments. We also include a focused characterization of the dataset via an Exploratory Data Analysis in the Methods section. 
653 |a Data analysis 
653 |a Data augmentation 
653 |a Accuracy 
653 |a Deep learning 
653 |a Datasets 
653 |a Classification 
653 |a Artificial intelligence 
653 |a Sentiment analysis 
653 |a Social networks 
653 |a Neural networks 
653 |a Ablation 
653 |a Natural language processing 
653 |a Multilingualism 
653 |a Robustness (mathematics) 
653 |a Machine learning 
653 |a Annotations 
653 |a Real time 
653 |a Emojis 
653 |a Digital media 
653 |a Semantics 
700 1 |a Liu, Yang  |u College of Arts & Sciences, University of Miami, Miami, FL 33124, USA; yxl2140@miami.edu 
700 1 |a Zhang, Tianhui  |u College of Engineering, Northeastern University, Boston, MA 02115, USA; zhang.tianhu@northeastern.edu 
700 1 |a Hou Lingmin  |u Department of Computer Science, Rochester Institute of Technology, Rochester, NY 14623, USA; lh1026@rit.edu 
700 1 |a Liu, Xianchen  |u Department of Computer Engineering, Florida International University, Miami, FL 33199, USA; xliu073@fiu.edu 
700 1 |a Guo Zhen  |u Department of Material Engineering, Florida International University, Miami, FL 33199, USA; zguo013@fiu.edu 
700 1 |a Mulati Aliya  |u Department of Politics and International Relations, Florida International University, Miami, FL 33199, USA; axm9395@miami.edu 
773 0 |t Systems  |g vol. 13, no. 11 (2025), p. 964-983 
786 0 |d ProQuest  |t Advanced Technologies & Aerospace Database 
856 4 1 |3 Citation/Abstract  |u https://www.proquest.com/docview/3275564978/abstract/embedded/7BTGNMKEMPT1V9Z2?source=fedsrch 
856 4 0 |3 Full Text + Graphics  |u https://www.proquest.com/docview/3275564978/fulltextwithgraphics/embedded/7BTGNMKEMPT1V9Z2?source=fedsrch 
856 4 0 |3 Full Text - PDF  |u https://www.proquest.com/docview/3275564978/fulltextPDF/embedded/7BTGNMKEMPT1V9Z2?source=fedsrch