赵蓉英+魏明坤

〔摘 要〕[目的]随着科学技术的不断发展，网络化发展的现象越来越受到人们的重视。如何在海量的网络信息中识别人物观点成为研究者关注的焦点，网络人物观点被视为网络文本表达的主要思想，是构成网络信息的“魂”。在海量的网络信息中快速识别网络人物观点对掌握网络信息主题具有重要作用。[方法]本文在前人研究的基础上理论与应用相结合，从文本分析的视角研究网络人物的观点。利用相应的算法对文本内容进行预处理，再通过文本句子中的词汇、词性标注和词汇之间的距离关系实现观点指示动词识别和观点持有者识别，从而实现网络人物的观点识别。[结果]通过网络人物观点算法识别的实证研究发现，通过对网络人物进行指代消解和观点持有者的扩展能有效地提高观点识别的准确率。

〔关键词〕文本分析；观点识别；观点持有者；观点指示动词；网络人物；网络信息

DOI：10.3969/j.issn.1008-0821.2017.12.017

〔中图分类号〕G250.252 〔文献标识码〕A 〔文章编号〕1008-0821（2017）12-0096-06

〔Abstract〕[Objective]With the development of science and technology，more and more attention has been paid to the phenomenon of network development.How to identify the views of people in the large information online has become the focus of attention of the researchers，the online peoples viewpoint is regarded as the main idea of the online text，which is the soul of the network information.It is very important to identify the viewpoint of characters online in the massive information.[Methods]This paper analyzed the viewpoint of network characters viewpoint from the perspective of text analysis on the basis of previous studies.This article was processed by using the corresponding algorithm，and then through the distance between the sentence words，lexical tagging and lexical relationship to achieve the identification of verbs and opinion holders，in order to realize the recognition on network peoples viewpoint.[Results]It realized recognizing the opinion of people in the network based on the previous research.The final results of experiment showed that the characters of opinion recognition would be effectively improved by eliminating the anaphora and expanding the opinion holders.

〔Key words〕text analysis；recognition of viewpoint；opinion holder；indicative verb of viewpoint；network character；network information

隨着Web2.0的出现，用户不仅是网络内容的浏览者，也是网络内容的发布者。人们除了可以从网络上获取更多的信息外，还可以通过网络评论网站、个人博客、社区、论坛等传递自己的观点。在此环境下，以网络评论为代表的主观性文本数量迅速增长，对网络人物观点的分析具有广泛的实用价值，如网络舆情分析、产品质量评论、影视和服务评价等。在这种背景下，以观点分析为主题的研究引起了学术界的广泛兴趣。……