| 新媒体时代网络评论观点信息发现机制研究 |
| |
| 引用本文： | 曾润喜,王君泽,杜洪涛.新媒体时代网络评论观点信息发现机制研究[J].图书情报工作,2015,59(14):111. |
| |
| 作者姓名： | 曾润喜 王君泽 杜洪涛 |
| |
| 作者单位： | 1. 重庆大学新闻学院 重庆 401331;
2. 华中科技大学公共管理学院 武汉 430074;
3. 清华大学公共管理学院 北京 100084 |
| |
| 基金项目： | 本文系中央高校基本科研业务费"基于大数据的社会舆情决策机制研究"(项目编号:106112015CDJSK07XK15)研究成果之一。 |
| |
| 摘 要： | 目的/意义] 舆情分析工作中的一项重要任务就是从冗杂的网络评论中发现具有代表性的网民观点信息。方法/过程] 提出一种无监督的信息发现方法,从网络评论内容中抽取具有代表性的网民观点表述。具体发现过程为,在网络评论中选取作为候选的观点表述语句,继而综合考察评论内容与相应文章的相关程度以及候选的观点表述语句中词汇之间的点互信息并对观点表述语句进行信息加工,最终得到具有代表性的用户观点表述信息。结果/结论] 本文所提出的信息发现方法取得了较好的实验检验效果,可为新媒体时代舆情研判工作者在繁杂冗余的网络评论观点中自动发现观点信息提供决策支持。
|
| 关 键 词： | 网络舆情 舆情分析 评论挖掘 观点抽取 |
| 收稿时间： | 2015-06-05 |
| 修稿时间： | 2015-07-04 |
Research on Opinion Extraction Mechanism of Web Information Reviews in New Media Times |
| |
| Zeng Runxi,Wang Junze,Du Hongtao.Research on Opinion Extraction Mechanism of Web Information Reviews in New Media Times[J].Library and Information Service,2015,59(14):111. |
| |
| Authors: | Zeng Runxi Wang Junze Du Hongtao |
| |
| Institution: | 1. School of Journalism and Communication, Chongqing University, Chongqing 401331;
2. School of Public Administration, Huazhong University of Sci&Tech, Wuhan 430074;
3. School of Public Policy and Management, Tsinghua University, Beijing 100084 |
| |
| Abstract: | Purpose/significance] Identifying the public typical opinions from the redundant web reviews is helpful for public opinion analysis task. Method/process] This paper proposes a method how to extract representative public opinions from web reviews with an unsupervised method. Firstly, it extracts the sentences from reviews which contains the high frequency terms reviews content, based on the template pre-defineded; secondly, it considers the relevance between reviews and corresponding articles, and extracts the pointwise-mutual information between terms in the sentences of candidate public opinions. Then it disposes the sentences and gets the final results that can reflect netizens' opinions.Result/conclusion] The experiment result shows that the method proposed can effectively identify the sentences that have informative, readable, and convey the major opinions. |
| |
| Keywords: | online public opinion public opinion analysis reviews mining opinion extraction |
|
| 点击此处可从《图书情报工作》浏览原始摘要信息 |
| 点击此处可从《图书情报工作》下载免费的PDF全文 |
|
