Network Public Opinion Tendency Detection Method Based on Sentiment Analysis

doi:10.11871/jfdc.issn.2096-742X.2026.01.008

Abstract

Abstract:

[Application Background] With the vigorous development of online social media, controversial online public opinions occur frequently, bringing numerous potential risks and challenges to society. As an effective approach for determining public opinion trends, sentiment analysis holds significant importance in the study of online public opinion. [Method] This paper presents an opinion tendency detection method based on BERT. This method devises a deep learning classifier for sentiment tendency analysis and, simultaneously, incorporates the contrastive learning strategy. [Conclusion] The proposed method deepens the model's representation of polarized disputes, remarkably improves the performance of negative public opinion tendency detection, and enhances the ability to detect public opinion polarization. This research offers a practical technical solution for detecting online public opinion risks, facilitating the timely identification and response to potential negative public opinions and polarized disputes, thus safeguarding the health and stability of the online environment.

Key words: sentiment analysis, deep learning, contrastive learning, public opinion analysis, polarized disputes

DENG Yiru,HE Hongbo,WANG Ying,WANG Runqiang. Network Public Opinion Tendency Detection Method Based on Sentiment Analysis[J]. Frontiers of Data and Computing, 2026, 8(1): 91-102, https://cstr.cn/32002.14.jfdc.CN10-1649/TP.2026.01.008.

Figures/Tables 8

Fig.1

Fig.2

Fig.3

Fig.4

Fig.5

Table 1

Results of the ablation experiment (%)"

method	$A C C n e g$	$A C C p o s$	$A C C a l l$	$F 1 m i c r o$
$B E R T b a s e l i n e$	81.02	81.30	81.23	72.05
$B E R T w_e m o M a s k$	82.57	82.35	82.44	72.98
$B E R T w_d e e p n e t$	83.24	83.01	83.12	73.84
$B E R T w_S u p C o n$	82.81	82.98	82.87	73.23
$B E R T w_a l l$	84.26	84.45	84.39	74.18

Table 1

Table 2

Fig.6

References 22

[1]	汤景泰, 徐铭亮, 星辰. 立场、情感、注意力与选择性接触: 舆论极化的影响要素分析[J]. 国际新闻界, 2023, 45(1): 132-156.
[2]	ZHANG X, HE Z, ZHAO R, et al. Research on Public Opinion Risk Management and Control Under Network Media Big Data[C]// EAI International Conference, BigIoT-EDU. Cham: Springer Nature Switzerland, 2022: 147-152.
[3]	SUHAIMIN M S M, HIJAZI M H A, MOUNG E G, et al. Social media sentiment analysis and opinion mining in public security: Taxonomy, trend analysis, issues and future directions[J]. Journal of King Saud University-Computer and Information Sciences, 2023, 35(9): 101776.
[4]	RODRíGUEZ-IBáNEZ M, CASáNEZ-VENTURA A, CASTEjóN-MATEOS F, et al. A review on sentiment analysis from social media platforms[J]. Expert Systems with Applications, 2023, 223: 119862.
[5]	JAIN R, KUMAR A, NAYYAR A, et al. Explaining sentiment analysis results on social media texts through visualization[J]. Multimedia Tools and Applications, 2023, 82(15): 22613-22629.
[6]	高浩鑫. 基于图神经网络的网络舆情预警风险智能监管[D]. 北京: 北京邮电大学, 2023.
[7]	管雨翔, 王娟, 兰月新, 等. 基于LDA和TextCNN的跨平台网络舆情风险预警研究[J]. 情报探索, 2024(10): 109-115.
[8]	罗涛, 谢凤祥, 李光华. 面向舆情监控的智能化自然语言处理算法设计[J]. 电子设计工程, 2023, 31(21): 114-118.
[9]	JEMAI F, HAYOUNI M, BACCAR S. Sentiment analysis using machine learning algorithms[C]// 2021 International Wireless Communications and Mobile Computing (IWCMC). IEEE, 2021: 775-779.
[10]	KOKAB S T, ASGHAR S, NAZ S. Transformer-based deep learning models for the sentiment analysis of social media data[J]. Array, 2022, 14: 100157.
[11]	佘梦婷. 基于BERT-BiLSTM-CNN多特征融合研究及在三胎政策舆情分析的应用[D]. 南昌: 南昌大学, 2023.
[12]	翟晗名. 基于知识图谱的社交媒体舆论热点问题研究[D]. 北京: 中国人民公安大学, 2024.
[13]	苏婕. 区域旅游舆情监测平台的设计与实现[D]. 成都: 电子科技大学, 2024.
[14]	邓任敏. 基于事理图谱的网络舆情分析与应用研究[D]. 天津: 天津科技大学, 2023.
[15]	JHA V, SAVITHA R, SHENOY P D, et al. A novel sentiment aware dictionary for multi-domain sentiment classification[J]. Computers & Electrical Engineering, 2018, 69: 585-597.
[16]	VIEGAS F, ALVIM M S, CANUTO S, ET AL. EXPLOITING SEMANTIC RELATIONSHIPS FOR UNSUPErvised expansion of sentiment lexicons[J]. Information Systems, 2020, 94: 101606.
[17]	GAMON M. Sentiment classification on customer feedback data: noisy data, large feature vectors, and the role of linguistic analysis[C]// COLING 2004: Proceedings of the 20th international conference on computational linguistics. 2004: 841-847.
[18]	YE Q, ZHANG Z, LAW R. Sentiment classification of online reviews to travel destinations by supervised machine learning approaches[J]. Expert systems with applications, 2009, 36(3): 6527-6535.
[19]	BUčAR J, ŽNIDARšIč M, POVH J. Annotated news corpora and a lexicon for sentiment analysis in Slovene[J]. Language Resources and Evaluation, 2018, 52: 895-919.
[20]	TIWARI P, MISHRA B K, KUMAR S, et al. Implementation of n-gram methodology for rotten tomatoes review dataset sentiment analysis[M]// Cognitive analytics: concepts, methodologies, tools, and applications. IGIGlobal, 2020: 689-701.
[21]	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, volume 1 (long and short papers). 2019: 4171-4186.
[22]	KHOSLA P, TETERWAK P, WANG C, et al. Supervised contrastive learning[J]. Advances in neural information processing systems, 2020, 33: 18661-18673.

Methods	Acc	micro_F1
CNN_attention	68.23	62.04
GRU_attention	77.62	67.88
BILSTM_attention	79.02	71.32
BIGRU	78.45	69.11
BIGRU_attention	79.87	70.87
BERT_w_all	84.39	74.18