[1]黄卫东,程小香.基于微博平台的舆情参与主体情感强度研究[J].计算机技术与发展,2022,32(11):140-145.[doi:10. 3969 / j. issn. 1673-629X. 2022. 11. 021]
　HUANG Wei-dong,CHENG Xiao-xiang.Research on Emotional Intensity of Public Opinion Participants Based on Microblog Platform[J].,2022,32(11):140-145.[doi:10. 3969 / j. issn. 1673-629X. 2022. 11. 021]
点击复制

基于微博平台的舆情参与主体情感强度研究()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
32
期数:
2022年11期
页码:
140-145
栏目:
人工智能
出版日期:
2022-11-10

文章信息/Info

Title:
Research on Emotional Intensity of Public Opinion Participants Based on Microblog Platform
文章编号:
1673-629X(2022)11-0140-06
作者:
黄卫东; 程小香
南京邮电大学 管理学院,江苏 南京 210003
Author(s):
HUANG Wei-dong; CHENG Xiao-xiang
School of Management,Nanjing University of Posts and Telecommunications,Nanjing 210003,China
关键词:
舆情; 参与主体; 情感强度; Bilstm; 注意力机制; 情感词典
Keywords:
public opinion; participants; emotional intensity; Bilstm; attention mechanism; emotional dictionary
分类号:
TP389. 1
DOI:
10. 3969 / j. issn. 1673-629X. 2022. 11. 021
摘要:
舆情参与主体是舆情事件的参与者,分析其情感强度可以准确把握舆情发展走势,为舆情相关部门的决策提供支持。 利用微博社交媒体平台获取实验所需的热点话题的舆情数据,基于 Word2vec 计算目标文本中关键词相似度并提取关键性特征。 将目标文本转化成向量形式嵌入到深度学习模型的输入层,同时将注意力机制引入深度学习算法构建Bilstm+Attention 情感倾向分类模型,对舆情参与主体发布在微博平台上文本的情感倾向进行正负面的划分。 为了进一步分析参与主体的情感强度以及与实际舆情走势之间的关系,基于目标语料构建情感副词词典,将情感强度进行划分,并与实际舆情走势进行对比分析。 实验结果表明, 对比 TextCNN、 CNN + Bilstm 以及 Bilstm 等深度学习分类模型, Bilstm +Attention 情感分类模型准确率更高,Bilstm+Attention+情感副词词典计算出的情感强度与实际舆情走势基本趋于一致,证明了该模型可以有效预测舆情参与主体的情感强度。
Abstract:
Public opinion participants are participants in public opinion events. Analyzing their emotional intensity can accurately graspthe development trend of public opinion and provide support for decision - making of relevant departments. Micro - blog social mediaplatform is used to obtain the public opinion data of hot topics needed in the experiment. Based on Word2vec,the similarity of keywordsin the target text is calculated and the key features are extracted. The target text is transformed into vector form and embedded into theinput layer of the deep learning model.? ? ?At the same time, the attention mechanism is introduced into the deep learning algorithm toconstruct the Bilstm + Attention sentiment orientation classification model, and the sentiment orientation of the text published on themicroblog platform by the public opinion participants is divided into positive and negative. In order to further analyze the relationshipbetween the emotional intensity of participants and the trend of actual public opinion,the emotional adverb dictionary is constructed basedon the target corpus,and the emotional intensity is divided and compared with the actual public opinion trend. The experimental resultsshow that compared with TextCNN,CNN+Bilstm and Bilstm,Bilstm+Attention has higher accuracy. The emotional intensity calculatedby Bilstm+Attention+ emotional adverb dictionary is basically consistent with the actual trend of public opinion,which proves that themodel can effectively predict the emotional intensity of public opinion participants.

相似文献/References:

[1]谢丽,丁海欣.基于 Bass 模型的谣言传播与控制问题研究[J].计算机技术与发展,2018,28(11):103.[doi:10.3969/ j. issn.1673-629X.2018.11.023]
　XIE Li,DING Hai-xin.Research on Propagation and Control of Rumor Based on Competitive Bass Model[J].,2018,28(11):103.[doi:10.3969/ j. issn.1673-629X.2018.11.023]
[2]毛建景,张君君.基于粒度商空间下的话题识别与跟踪研究[J].计算机技术与发展,2019,29(07):190.[doi:10. 3969 / j. issn. 1673-629X. 2019. 07. 038]
　MAO Jian-jing,ZHANG Jun-jun.Research on Topic Recognition and Tracking Based on Granular Quotient Space[J].,2019,29(11):190.[doi:10. 3969 / j. issn. 1673-629X. 2019. 07. 038]

更新日期/Last Update: 2022-11-10