Article Details

Abstract

With the full advent of the We-Media era, public opinion governance in colleges and universities becomes an important issue to be addressed urgently. As a major social media platform, Weibo becomes the core vehicle for the dissemination of public opinion in colleges and universities. Taking the Weibo platform as a case study, this paper independently constructs a corpus of university-related Weibo posts and an internet slang sentiment lexicon, and proposes a BERT-BiLSTM-Attention sentiment classification model fused with features derived from the internet slang sentiment lexicon. Latent Dirichlet Allocation (LDA) is employed for topic mining to implement a multi-dimensional analysis covering the sentiment, topic, and time dimensions. Based on the multi-dimensional analytical results, differentiated public opinion guidance strategies are proposed. Experimental results demonstrate that the proposed model achieves favorable performance in terms of accuracy, precision, recall, and F1-score. The proposed guidance strategies provide effective guidance for the direction of public opinion governance in colleges and universities.

Keywords

university network public opinion; Deep Learning; sentiment classification model; LDA; multi-dimensional analysis

References

1. [1] China Internet Network Information Center The 56th Statistical Report on Internet Development in China [R]. 2025-07-21
2. [2] Sina Technology Weibo releases Q3 financial report with total revenue of 3.156 billion yuan [EB/OL]. (2025-11-18) [2026-01-12] https://weibo.com/ttarticle/p/show?id = 2309405234379378524275.
3. [3] Li Botao, Xi Hongqi, Chen Danmin A university public opinion prediction system based on sentiment analysis [J]. Journal of Henan University of Finance and Economics (Natural Science Edition), 2024, 33 (3): 14-19
4. [4] Huang Ping, Zhu Huijuan, Chen Linlin Research on the Application of Emotion Classification Technology Based on Deep Learning in University Public Opinion Analysis [J]. Software Engineering, 2021, 24 (11): 59-62
5. [5] Zhang Qian Monitoring and Analysis of University Public Opinion Based on Weibo [J]. Modern Information Technology, 2025, 9 (20): 75-78
6. [6] Devlin J，Chang M-W，Lee K，et al. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding [C]//Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies，2019（1）：4171-4186.
7. [7] Wang Xue, Zhang Shuisheng, Wei Yunbo Empirical analysis of online public opinion in universities based on BERT-LSTM model [J]. Journal of Qiqihar University (Natural Science Edition), 2024, 40 (5): 87-90
8. [8] Sun Z J，Li X Y，Sun X F，et al. ChineseBERT: Chinese Pretraining Enhanced by Glyph and Pinyin Information [C]//Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing，2021（1）：2065-2075.
9. [9] Hochreiter S，Schmidhuber J. Long Short-Term Memory [J].Neural Computation，1997，9（8）：1735-1780.
10. [10] Vaswani A，Shazeer N，Parmar N，et al. Attention Is All You Need [J/OL].arXiv:1706.03762 [cs.CL].https://arxiv.org/abs/1706.03762.
11. [11] Zeng Ziming, Chen Siyu Evolutionary analysis of network public opinion on sudden public health emergencies based on LDA and BERT BiLSTM Attention models [J]. Intelligence Theory and Practice, 2023, 46 (9): 158-166

Funding

School level Project of Inner Mongolia Electronic Information Vocational and Technical College (KZY2025009)

About the authors

Shi Na (1999-), female, Han ethnicity, from Ulanqab, Inner Mongolia, holds a Master's degree in Engineering with research interests in artificial intelligence and natural language processing.