Article Details

Abstract

Against the backdrop of advancing globalization, social media public opinion in Northeast Asia exerts an increasingly significant influence on the region's politics, economy, culture, and other aspects. This study aims to develop a social media public opinion analysis method tailored to Northeast Asia. Through the collection and integration of hot-topic public opinion data from major social media platforms in the region, a Northeast Asia public opinion dataset comprising 40 610 texts across nine categories is constructed. The BERTopic model is employed to extract key topics and trending topics within each category, while Baidu AI sentiment analysis is utilized to mine public emotional tendencies. Experimental results show that the proposed method achieves a topic coherence score of 0.62, outperforming traditional LDA and NMF topic models. The Baidu AI engine also attains a Macro-F1 score of 90.5% in sentiment analysis tasks. The two-dimensional cross-analysis of topic and sentiment provides a quantifiable technical pathway for assessing regional public opinion situations.

Keywords

Northeast Asia; social media; public opinion analysis; sentiment analysis; BERTopic

References

1. [1] Zhang Yunling, Yang Bojiang, Xiang Haoyu Review and Prospect of the Situation in Northeast Asia under the Deep Evolution of a Century of Changes [J]. Northeast Asia Journal, 2024 (2): 1-13143
2. [2] Kui Hao, Diao Hongyue, Kong Liangchen, etc Research on Fine grained Named Entity Recognition Method for Northeast Asian Public Opinion Text [J]. Computer Engineering, 2024, 50 (5): 354-362
3. [3] Grootendorst M. BERTopic: Neural Topic Modeling with a Class-based TF-IDF Procedure [PP/OL].arXiv:2203.05794 [cs.CL].https://arxiv.org/abs/2203.05794.
4. [4] Shi Wei, Xue Guangcong, He Shaoyi A review of research on online public opinion from an emotional perspective [J]. Journal of Library and Information Science, 2022, 39 (1): 105-118
5. [5] Pennebaker J W，Francis M E，Booth R J. Linguistic Inquiry and Word Count: LIWC 2001 [M].Mahwah：Lawrence Erlbaum Associates，2001.
6. [6] Wilson T，Wiebe J，Hoffmann P. Recognizing Contextual Polarity: An Exploration of Features for Phrase-level Sentiment Analysis [J].Computational Linguistics，2009，35（3）：399-433.
7. [7] Han Kun, Pan Hongpeng, Liu Zhongyi Research on sentiment analysis of short video network public opinion by integrating BERT multi-level features [J]. Computer Science and Exploration, 2024, 18 (4): 1010-1020
8. [8] Li J，Sun A X，Han J L，et al. A Survey on Deep Learning for Named Entity Recognition [J].IEEE Transactions on Knowledge and Data Engineering，2022，34（1）：50-70.
9. [9] An Y，Oh H，Lee J. Marketing Insights from Reviews Using Topic Modeling with BERTopic and Deep Clustering Network [J].Applied Sciences，2023，13（16）：9443.
10. [10] Hutama L B，Suhartono D. Indonesian Hoax News Classification with Multilingual Transformer Model and BERTopic [J].Informatica，2022，46（8）：81-90.
11. [11] Zhou Hong, Yu Hailong, Wang Liping, etc Research on cutting-edge monitoring and thematic analysis of lithium batteries based on BERTopic topic model [J]. Energy Storage Science and Technology, 2025, 14 (1): 406-416

Funding

Basic Research Project of Higher Education Institutions in Liaoning Province, China (LJKQZ20222451)

About the authors

Ji Lupeng (1989-), male, Han ethnicity, from Heze, Shandong, engineer, bachelor's degree, research interests include data mining; Corresponding author: Wei Hao (1993-), male, Han ethnicity, from Jinan, Shandong, associate professor, master's supervisor, doctoral student, research interests include natural language processing; Diao Hongyue (1992-), female, Han ethnicity, from Chaoyang, Liaoning, lecturer, master's supervisor, doctoral student, research interests include intelligent information processing.