| 摘 要: 针对传统评论方式依赖整体感知且相对滞后的问题，以弹幕这一新兴短信息表达方式为研究对象，采用文本挖掘与情感分析的方式研究弹幕与网络舆情之间的潜在联系。采用网络爬虫技术采集网络舆情弹幕数据，使用Jieba库实现分词、去停用词及高频词统计，基于WordCloud库绘制词云图，实现可视化，并使用SnowNLP库计算网络舆情弹幕的情感得分，运用隐含狄利克雷分布(Latent Dirichlet Allocation, LDA)模型进行主题词提取，实现对网络舆情弹幕的情感分类和主题分析。实验结果表明，该方法可多维展现网民的情感倾向与关注焦点，是对传统评论文本研究的有效补充。 |
| 关键词: 网络舆情；情感分析；文本挖掘；SnowNLP；LDA |
|
中图分类号: TP391.1
文献标识码: A
|
|
| Text Mining and Sentiment Analysis of Network Public Opinion based on Bullet Screen |
|
BAI Jian, HONG Xiaojuan
|
(School of Management, Nanjing University of Posts and Telecommunications, Nanjing 210003, China)
1535179246@qq.com; 1291823970@qq.com
|
| Abstract: In view of the problem that the traditional review methods rely on the overall perception and are relatively lagging behind, this paper takes the bullet screen, a new short message expression, as the research object, and proposes to use text mining and sentiment analysis to further study the potential relationship between bullet screen and network public opinion. Firstly, network crawler technology is used to collect the bulletin data of network public opinion. Secondly, Jieba library is used to realize word segmentation, stop words and high frequency word statistics, and WordCloud library is used to draw word cloud map to realize visualization. Finally, SnowNLP library is used to calculate the sentiment score of the network public opinion bullet screen, and LDA (Latent Dirichlet Allocation) model is used to extract the keywords to realize the sentiment classification and theme analysis of the network public opinion bullet screen. The experimental results show that this method can show the sentiment tendency and focus of netizens in multiple dimensions, and it is an effective supplement to traditional review text research. |
| Keywords: network public opinion; sentiment analysis; text mining; SnowNLP; LDA |
