Sentiment analysis of microblog public opinion based on deep learning

Yi  Xie; Kaiyu  Li; Zhaomin  Liu; Shicheng  Li; Di  Liu

doi:10.54097/apdrq576

Authors

Yi Xie
Kaiyu Li
Zhaomin Liu
Shicheng Li
Di Liu

DOI:

https://doi.org/10.54097/apdrq576

Keywords:

microblog public opinion analysis; deep learning; BERT model; Attention mechanism; GRU

Abstract

Public opinion sentiment analysis focuses on the public's emotional bias towards public events, and public opinion involving public health events will directly affect social stability, so it is essential for microblog sentiment analysis. In this paper, we use the RoBERTa-BiGRU-Attention model to analyze the sentiment of the text data of Weibo by taking the 2024 Paris Olympics as an example. Firstly, RoBERTa obtains the word embedding representation containing the text context information from the text data in the input layer. BiGRU obtains the character representation and sentence representation, and finally, the weight proportion of each character to the sentence in which it is located is calculated using the Attention mechanism, The text representation of the full text is obtained, and the softmax function carries out the sentiment analysis. To verify the effectiveness of the RoBERTa-BiGRU-Attention model, this paper takes Accuracy, Precision, Recall, and F1-measure as the evaluation indexes, and finds five models to compare with them. The results show that the RoBERTa-BiGRU-Attention model has good performance and advantages compared with the general model.

References

[1] Wu Jiahui, Jia Yungang, Wang Zhixiao, et al. Sentiment analysis of Weibo epidemic public opinion text based on deep learning[J].Computer Technology and Development, 024,34(07):175-183.DOI:10.20165/j.cnki.ISSN1673-629X.2024.0087.

[2] Xi Yuyi,Lv Guo,Bai Mohan. Research on microblog public opinion analysis based on CNN-BiLstm model[J].Journal of Hebei University of Civil Engineering and Architecture, 023,41(03):207-210+220.)

[3] ZHANG Li,LI Ju. Research on microblog public opinion sentiment analysis based on dictionary and emoji[J].Computer & Telecommunications,2023,(07):40-44.DOI:10. 5966/j. nki. nydx.2023.07.011.

[4] GUO Junhao. Research and application of text sentiment analysis based on deep learning[D].North University of China,2023.DOI:10.27470/d.cnki.ghbgc.2023.000118.

[5] ZHANG Li. Research on Chinese implicit sentiment analysis for social network text[D].East China University of Technology, 023.DOI:10.27145/d.cnki.ghddc.2023.000069.

[6] MA Yu. Research on microblog public opinion sentiment analysis and evolution trend prediction technology under major emergencies[D].Zhengzhou University of Light Industry, 023.DOI:10.27469/d.cnki.gzzqc.2023.000173.

[7] Li Yunxiang. Research on Chinese named entity recognition based on RoBERTa-WWM[D].Nanning Normal University, 023.DOI:10.27037/d.cnki.ggxsc.2023.000246.

[8] Wang Shuyan,Yuan Ke. Sentiment analysis model of college student forum based on RoBERTa-WWM[J].Computer Engineering, 022,48(08):292-298+ 05.DOI: 0.19678/j. ssn. 000- 428.0062008.

[9] LIU Siyu. Research on improvement of Chinese named entity recognition method based on deep learning[D].Chengdu University of Technology, 020.DOI:10.26986/d. nki. cdlc. 020.000601.)