Article Details

Abstract

With the rapid development of social media, Weibo, as an important platform for user information exchange and emotion expression, accumulates massive and rich text data. Based on the investigation of related technologies for Weibo sentiment analysis, this paper proposes a fine-grained sentiment analysis method based on the BERT model. Combining Weibo data crawling and preprocessing technologies, this paper constructs an efficient analysis framework. This method completes data collection through the Weibo Open API, utilizes the BERT pre-training model to realize text vectorization, and completes the precise classification of six emotions including anger, happiness, neutral, surprise, sadness, and fear based on the Transformer architecture. Meanwhile, this paper introduces sentiment dictionaries and data augmentation technologies to improve model performance, and displays analysis results by means of visualization tools. The study shows that this method achieves high classification accuracy on the SMP2020 dataset and provides a new idea for fine-grained sentiment analysis of Weibo text.

Keywords

Weibo sentiment analysis; fine-grained sentiment; BERT model; social media

References

1. [1] Lang Cong Deep Learning based Sentiment Analysis of Weibo Text [D]. Shenyang: Shenyang University of Technology, 2021
2. [2] Cai Zengyu, Han Yang, Zhang Jianwei, etc SnowNLP based Weibo Network Public Opinion Analysis System [J]. Science, Technology and Engineering, 2024, 24 (13): 5457-5464
3. [3] Willow trees Algorithm Analysis and System Implementation of Weibo Public Opinion [D]. Dalian: Dalian University of Technology, 2020
4. [4] Zhu He, Lu Xiaofeng, Xue Lei A Financial Text Sentiment Analysis Model Based on BERT [J]. Journal of Shanghai University: Natural Science Edition, 2023, 29 (1): 118-128
5. [5] Han Guang, Gao Ming, Zhong Jiachen Research on Sentiment Analysis of Aviation Review Data Based on BERT Model [J]. Microcomputer Applications, 2024, 40 (6): 250-252
6. [6] LI Y. A Knowledge Enhanced Pre-Training Model for Chinese Weibo Sentiment Analysis [C]//2024 9th International Conference on Cloud Computing and Big Data Analytics (ICCCBDA).Chengdu：IEEE，2024：33-39.
7. [7] XIONG J J，LIU W，MA J W. Dissemination and Sentiment Analysis of Mega Sports Events Based on Big Data Statistics Under Social Media Environment [J].Journal of Systems Science and Information，2025，13（2）：221-239.
8. [8] Guo Xiaolei, Zhang Wubo Text classification of citizen hotlines based on short text extension and feature fusion [J]. Journal of Shanxi Datong University: Natural Science Edition, 2024, 40 (1): 42-47, 62
9. [9] Yang Yufan, Gao Dingguo, Xu Zezhou Research on Text Sentiment Classification Method Based on Transformer [J]. Journal of Central University for Nationalities: Natural Science Edition, 2022, 31 (3): 42-45
10. [10] Zhao Shaoya Research on sentiment analysis and topic discovery of online public opinion in universities: based on data from a certain university's Weibo super talk community [J]. Science and Technology Communication, 2024, 16 (17): 91-95
11. [11] Yu Tongrui, Jin Ran, Han Xiaozhen, etc A review of research on pre trained models for natural language processing [J]. Computer Engineering and Applications, 2020, 56 (23): 12-22
12. [12] Wu Jiahui, Jia Yungang, Wang Zhixiao, etc Deep Learning based Sentiment Analysis of Weibo Epidemic Public Opinion Text [J]. Computer Technology and Development, 2024, 34 (7): 175-183
13. [13] Xi Yuyi Research on Weibo Public Opinion Analysis Based on Deep Learning [D]. Zhangjiakou: Hebei University of Architecture and Technology, 2024
14. [14] Zhang Zeyuan, Zhang Guangda, Li Jiayu, etc Daily Text Sentiment Analysis Based on BERT [J]. Journal of Qiqihar University: Natural Science Edition, 2024, 40 (6): 37-41
15. [15] Meng Haoyang Research on the Application of Deep Learning based Sentiment Analysis in Weibo Comments [D]. Shijiazhuang: Hebei University of Science and Technology, 2023

Funding

Shanghai Sanda University Key Course Project (A020201.24.049); Shanghai Sanda University Research Fund Project - Research on Solving Schedule Scheduling Problems Based on Reinforcement Learning Collaborative Evolutionary Algorithm

About the authors

Zhang Yimin (2002-), male, Han ethnicity, from Liyang, Jiangsu Province, undergraduate studies, research direction: natural language processing; Corresponding author: Li Ye (1981-), male, Han ethnicity, born in Hulunbuir, Inner Mongolia, lecturer, system analyst, master's degree, research direction: computational intelligence, distributed systems.