摘要
[目的/意义]及时准确地检测中亚国家“一带一路”网络舆情热点,对于推进“一带一路”建设具有积极意义。[方法/过程]首先利用ETM(Embedded Topic Model)模型对文档建模得到候选热点主题;然后借助YAKE关键词抽取算法综合考虑候选热点主题的一致性、独立性和媒体关注度,确定热点主题;最后抽取每一主题的代表性文档以增强热点主题的可理解性。[结果/结论]在构建的中亚国家“一带一路”网络舆情数据集上的实验表明,该方法取得了较好效果。
[Purpose/Significance]The timely and accurate detection of hot spots of"Belt and Road"network public opinion in Central Asian countries is of positive significance to China's promotion of the“Belt and Road”construction.[Method/Process]The paper first uses the ETM(embedded topic model)to model the document so as to obtain candidate hot topics;then it uses the YAKE keyword extraction algorithm to comprehensively evaluate the consistency,independence and media attention of the candidate hot topics,thus determining the hot topics;and finally it extracts representative documents of each hot topic to enhance the intelligibility of these hot topics.[Result/Conclusion]Experiments on the"Belt and Road"network public opinion data set in Central Asian countries indicate that the research method of this paper has achieved good results.
作者
席耀一
高鑫
王小明
云建飞
Xi Yaoyi;Gao Xin;Wang Xiaoming;Yun Jianfei(Information Engineering University,Zhengzhou 450001;Information Engineering University,Luoyang Campus,Luoyang 471003)
出处
《情报杂志》
CSSCI
北大核心
2020年第11期82-89,共8页
Journal of Intelligence
基金
国家社会科学基金青年项目“基于大数据技术的中亚国家‘一带一路’网络舆情研究”(编号:19CXW027)。
关键词
网络舆情
热点检测
热度评分
嵌入式主题模型
“一带一路”
中亚
network public opinion
Hot-Spot Detection
heat degree
embedded topic model
"Belt And Road"
Central Asia
