第三届中文倾向性分析评测(COAE2011)语料的构建与分析

文本倾向性分析已成为自然语言处理领域研究的热点问题之一。为进一步推动中文倾向性分析的研究,中国中文信息学会信息检索专业委员会举办了第三届中文倾向性分析评测(COAE2011)。该次评测主要关注领域和上下文语境(Context)对中文倾向性分析的影响。该文主要介绍COAE2011评测语料的构建及其对评测的支撑 首先介绍了COAE2011语料的领域选取、媒介分布等获取过程,然后详细阐述语料的标注原则与方法,最后依据评测结果分析领域和上下文语境因素对倾向性的影响。COAE2011语料的建立将为中文倾向性分析提供强大的资源支持。     

Opinion mining and sentiment analysis for Arabic on-line texts: application on the political domain

Since the continuous proliferation of the journalistic content online and the changing political landscape in many Arabic countries, we started our current research in order to implement a media monitoring system about the opinion mining in political field. This system allows political actors, despite of the large volume of online data, to be constantly informed about opinions expressed on the web in order to properly monitor their actual standing, orient their communication strategy and prepare the election campaigns. The developed system is based on a linguistic approach using NooJ’s linguistic engine to formalize the automatic recognition rules and apply them to a dynamic corpus composed of journalistic articles. The first implemented rules allow identifying and annotating the different political entities (political actors and organizations). Then these annotations are used in our system of media monitoring in order to identify the opinions associated with the extracted named entities. The system is mainly based on a set of local grammars developed for the identification of different structures of the political opinion phrases. These grammars are using the entries of the opinion lexicon that contain the different opinion words (verbs, adjectives, nouns) where each entry is associated with the corresponding semantic marker (polarity and intensity). Our developed system is able to identify and properly annotate the opinion holder, the opinion target and the polarity (positive or negative) of the phraseological expression (nominal or verbal) expressing the opinion. Our experiments showed that the adopted method of extraction is consistent with 0.83 F-measure.     

面向社会媒体的信息抽取：任务、挑战与机遇

社会媒体是人们用来分享意见、见解、观念和经验的平台或工具,目前已经发展成具有日益重大影响力的新媒体。面向社会媒体的信息抽取就是要从充满噪音的、零碎的、非结构化的社会媒体的自由文本中提取有价值的结构化的信息,以利于从社会媒体内容中有效地获取信息。介绍了面向社会媒体的信息抽取这一任务的诞生背景、要解决的主要问题、面l临的主要挑战、相关工作以及未来的研究方向。     

AI, E-government, and Politics 2.0

This issue's Trends and Controversies department includes five essays on e-government and politics 2.0 from distinguished experts. Each essay presents a unique, innovative research framework, computational methods, and selected results and examples. As the government and political process become more transparent, participatory, online, and multimedia rich, there is a great opportunity for adopting advanced AI and intelligent systems research in e-government and politics 2.0 applications. Selected techniques in data, text, Web, and opinion mining, social network analysis, visual analytics, multimedia analysis, ontological representations, and social media analysis can support online political participation, e-democracy, political blogs and forums, e-government service delivery, and transparency and accountability.     

Social Media Analysis and Public Opinion: The 2010 UK General Election

Social media monitoring in politics can be understood by situating it in theories of public opinion. The multimethod study we present here indicates how social media monitoring allow for analysis of social dynamics through which opinions form and shift. Analysis of media coverage from the 2010 UK General Election demonstrates that social media are now being equated with public opinion by political journalists. We use interviews with pollsters, social media researchers and journalists to examine the perceived link between social media and public opinion. In light of competing understandings these interviews reveal, we argue for a broadening of the definition of public opinion to include its social dimension.     

Media Reporting of Nuclear Emergencies: The Effects of Transparent Communication in a Minor Nuclear Event

IN 2008, a nuclear event occurred at Kr?ko nuclear power plant in Slovenia. Even though it was classified as level 0 on International Nuclear Event Scale, the transparency policy of the Slovenian nuclear safety authorities prompted it to notify the international community. This was the first time that the European Community Urgent Radiological Information Exchange (ECURIE) notification system was used outside the exercise framework. The event was reported in all major European. In this contribution, we report on a content analysis of media articles related to this event. The main research question was if a nuclear emergency generates significant media coverage, even in the case of a minor event and a transparent communication policy. The analysis included more than 200 articles from printed and spoken media in Slovenia and other countries. The analysis revealed a high‐intensity media coverage, emotional reactions and heated political discussion. The main media sources in countries with open political discussions on nuclear energy end up being the politicians, rather than resident experts.     

A survey on sentiment analysis and opinion mining for social multimedia

Social media sentiment analysis (also known as opinion mining) which aims to extract people’s opinions, attitudes and emotions from social networks has become a research hotspot. Conventional sentiment analysis concentrates primarily on the textual content. However, multimedia sentiment analysis has begun to receive attention since visual content such as images and videos is becoming a new medium for self-expression in social networks. In order to provide a reference for the researchers in this active area, we give an overview of this topic and describe the algorithms of sentiment analysis and opinion mining for social multimedia. Having conducted a brief review on textual sentiment analysis for social media, we present a comprehensive survey of visual sentiment analysis on the basis of a thorough investigation of the existing literature. We further give a summary of existing studies on multimodal sentiment analysis which combines multiple media channels. We finally summarize the existing benchmark datasets in this area, and discuss the future research trends and potential directions for multimedia sentiment analysis. This survey covers 100 articles during 2008–2018 and categorizes existing studies according to the approaches they adopt.

     

基于直推式学习的中文情感词极性判别

态度挖掘是近年来文本挖掘领域的热点课题之一,旨在发现文本中作者的主观态度倾向,为基于舆情的决策过程提供支持。目前已有的态度挖掘算法绝大多数都基于情感词典来识别情感词,在此基础上判别句子或文本的总体态度倾向。然而,手工构造和维护一部完善的情感词典是不现实的。对中文情感词的极性判别问题进行了研究,提出了基于直推式学习的中文情感词极性判别算法。该算法以少量情感词为种子,利用词典中词汇的解释信息,直推出其他词的情感极性。与使用相同情感种子词的解释信息作为训练数据的有监督学习算法相比,直推式学习算法的识别精度提高了20%左右。     

社会媒体短文本内容的语义概念关联和扩展

随着微博、照片分享等社会化媒体的快速发展,每天产生了大量的短文本内容如评论、微博等,对其进行深入挖掘有重大的应用价值和学术意义。该文选取微博作为例子,详细阐述我们提出的方法。微博信息流因其简短和实时的特性而具有非常大的价值,已经成为市场营销,股票预测、舆情监控等应用的重要信息源。尽管如此,微博内容特征极其稀疏、上下文语境提取困难,使得微博信息的挖掘面临着很大挑战。因此,我们提出一种基于Wikipedia的微博语义概念扩展方法,通过自动识别那些与微博信息语义相关的Wikipedia概念来丰富它的内容特征,从而有效提高微博信息数据挖掘和分析的效果。该文工作首先通过可链接性剪枝、概念关联和消歧,发现微博信息中重要的n-gram所对应的Wikipedia概念;其次,采用基于概念-文档关联矩阵的NMF分解(非负矩阵分解)方法获取Wikipedia概念之间的语义近邻,为微博信息扩展相关的语义概念。基于TREC 2011的微博数据集和Wikipedia 2011数据集进行实验,与已有两个相关研究工作比较,该文提出的方法取得了较好的效果。     

汉语意见型主观性文本类型体系的研究

主观性文本是一种描述个人想法、情感和意见等的非约束性文本。它与主要描述以事实为主的客观性文本在内容和结构上有很大的不同。意见型文本是包含有意见元素(意见持有者、意见陈述范围、意见主题和意见情感)的一种主观性文本,它大量出现在网上的电子公告板、论坛和博客等媒介中,受到广泛的关注,并成为研究意见挖掘方法和技术的语料。该文介绍了主观性文本的定义及其与客观性文本的差异,同时着重讨论了意见型文本的定义、特点、类型体系及其在意见挖掘技术中的应用。     

A systematic study on the role of SentiWordNet in opinion mining

Sentiment lexicons (SL) (aka lexical resources) are the repositories of one or several dictionaries that consist of known and precompiled sentiment terms. These lexicons play an important role in performing several different opinion mining tasks. The efficacy of the lexicon-based approaches in performing opinion mining (OM) tasks solely depends on selecting an appropriate opinion lexicon to analyze the text. Therefore, one has to explore the available sentiment lexicons and then select the most suitable resource. Among available resources, SentiWordNet (SWN) is the most widely used lexicon to perform tasks related to opinion mining. In SWN, each synset of WordNet is being assigned the three sentiment numerical scores; positive, negative and objective that are calculated using by a set of classifiers. In this paper, a detailed and comprehensive review of the work related to opinion mining using SentiWordNet is provided in a very distinctive way. This survey will be useful for the researchers contributing to the field of opinion mining. Following features make our contribution worthwhile and unique among the reviews of similar kind: (i) our review classifies the existing literature with respect to opinion mining tasks and subtasks (ii) it covers a very different outlook of the opinion mining field by providing in-depth discussions of the existing works at different granularity levels (word, sentences, document, aspect, clause, and concept levels) (iii) this state-ofart review covers each article in the following dimensions: the designated task performed, granularity level of the task completed, results obtained, and feature dimensions, and (iv) lastly it concludes the summary of the related articles according to the granularity levels, publishing years, related tasks (or subtasks), and types of classifiers used. In the end, major challenges and tasks related to lexicon-based approaches towards opinion mining are also discussed.     

Textual factors in online product reviews: a foundation for a more influential approach to opinion mining

While data mining is well established in practice, opinion mining is still in its infancy, with issues in particular around the development of methodologies which effectively extract accurate, reliable, influential and useful information from the raw opinion data collected from informal product reviews. Current approaches adopt a single-variable approach, focusing on individual metrics—word length, the presence of keywords, or the overall semantic orientation of terms within the data—while neglecting to evaluate whether these individual artifacts are indicative of the tone of a given review. This approach has significant limitations when we move from trying to merely evaluate whether an online opinion is positive or negative, to trying to evaluate how likely it is that the opinion will influence others. Given this issue, one promising avenue would be to evaluate the general analysis approaches utilized by opinion mining algorithms and identified in the literature in terms of how accurately they reflect how people actually interpret and are influenced by electronic online reviews. Through interviewing and a follow up survey of 136?participants, the validity of the approach in terms of ascertaining the tone of a piece of text can be evaluated, as well as the identification of measurable factors within text which make a given opinionated text more or less influential in an online context, further facilitating the development of more effective multivariate opinion mining approaches. Furthermore, the identification of factors which make an online opinion text more or less persuasive helps to facilitate the development of opinion mining approaches which can evaluate how likely a review is to affect an individual’s decision making.     

The Politics of Hostage Rescue: Is Violence a Route to Political Success?

Over the years, law enforcement agencies have acquired extensive experience with hostage incidents, and most Western countries have officers trained in all aspects of hostage resolution. There are also articles and manuals outlining how to deal with the media coverage of hostage takings (Scanlon, 1989). However, because hostage rescue efforts can provide dramatic visuals that attract enormous audiences, the media have steadily intensified their coverage of such incidents. Today, a group of previously obscure persons can suddenly dominate the media agenda by successfully resisting an armed assault or by seizing hostages and calling themselves terrorists. After defining a hostage incident and looking at the strategy for dealing with such incidents, this article examines the implications of two fatal incidents: the stand‐off involving religious fanatics at Waco, Texas; and the Air France hijacking that started in Algiers and ended in Marseille, France. Both became number one on the Western media agenda, and both became political crises involving the head of state; one threatening a president’s credibility, the other enhancing a president’s status. Together they suggest that the escalating media coverage of such incidents raises questions not only about the effectiveness of current response strategies, but also about political leadership. This article discusses a number of strategies that have been tried or suggested. It also debates whether involvement has a positive or negative effect on political leaders. It concludes that, from the evidence available, a successful hostage rescue can yield political rewards.     

一种基于OCC模型的文本情感挖掘方法

观点挖掘（或情感分析）作为面向网络社会媒体分析挖掘领域的一个核心研究课题,具有重要的研究意义和应用价值。针对传统观点挖掘方法存在的不足和局限性,本文设计并实现了一种基于OCC情感模型的观点挖掘方法。该方法首先采用统计方法,利用WordNet词典、句法依存关系及少量标注数据,自动构建情感维度词典;其次,对所构建的情感维度词典进行求精,通过语义、情感倾向的不一致性处理和非情感词的过滤,得到高质量的情感维度词典;最后,基于所得到的情感维度词典,结合OCC模型中情感维度值与情感类型的对应关系,生成6种主要的情感类型。实验方法表明,此方法在使用灵活性、可解释性和有效性上具有明显的优势。     

Perceptions of News Credibility about the War in Iraq: Why War Opponents Perceived the Internet as the Most Credible Medium

This study investigated cross‐media credibility perception with respect to news coverage about the Iraq War. In an environment of political partisanship, perceptions of media credibility were likely affected by the audience’s political position on the war. Based on hostile media effect theory, a set of hypotheses was proposed to investigate whether the minority opinion group, war opponents, evaluated the Internet as a more credible medium than did neutrals or supporters. An online survey was conducted to which 481 people responded (71% war supporters, 19% opponents, 10% neutrals). Results showed that opponents of the war perceived the Internet as less aligned with a pro‐government position and as more credible than did neutrals or supporters. The opponent group also showed a strong negative correlation between perceived pro‐government alignment and perceptions of Internet credibility. For the minority partisan group, the diversity of information and views on the war was the main reason for the perception of high credibility of the Internet as a news channel.     

User opinion classification in social media: A global consistency maximization approach

Social media is a major platform for opinion sharing. In order to better understand and exploit opinions on social media, we aim to classify users with opposite opinions on a topic for decision support. Rather than mining text content, we introduce a link-based classification model, named global consistency maximization (GCM) that partitions a social network into two classes of users with opposite opinions. Experiments on a Twitter data set show that: (1) our global approach achieves higher accuracy than two baseline approaches and (2) link-based classifiers are more robust to small training samples if selected properly.     

WSAM:互联网UGC文本主观观点挖掘系统

互联网上的用户生成内容UGC(User Generated Content)中蕴含的用户主观观点信息对分析用户行为、用户需求等工作有着重要的价值。设计一套基于自然语言理解的互联网UGC文本主观观点分析系统WSAM,该系统能挖掘出用户主观观点所蕴含的关注对象和主观成分。分析了互联网UGC现象和生成原因,总结出UGC中用户主观观点中的四种主要类型。挖掘用户主观观点过程中,将用户主观观点的挖掘转化为句子中主观观点关注对象的识别和主观成分的判断。算法结合基于词语类、结构类等相关特征,采用最大熵分类器挖掘用户主观观点。实验验证,WSAM系统所采用的算法性能较好,且还能够灵活扩充出情感分析(Opin-ion Mining)等相关应用,同样也能达到较好的结果。     

Opinion mining from noisy text data

The proliferation of Internet has not only led to the generation of huge volumes of unstructured information in the form of web documents, but a large amount of text is also generated in the form of emails, blogs, and feedbacks, etc. The data generated from online communication acts as potential gold mines for discovering knowledge, particularly for market researchers. Text analytics has matured and is being successfully employed to mine important information from unstructured text documents. The chief bottleneck for designing text mining systems for handling blogs arise from the fact that online communication text data are often noisy. These texts are informally written. They suffer from spelling mistakes, grammatical errors, improper punctuation and irrational capitalization. This paper focuses on opinion extraction from noisy text data. It is aimed at extracting and consolidating opinions of customers from blogs and feedbacks, at multiple levels of granularity. We have proposed a framework in which these texts are first cleaned using domain knowledge and then subjected to mining. Ours is a semi-automated approach, in which the system aids in the process of knowledge assimilation for knowledge-base building and also performs the analytics. Domain experts ratify the knowledge base and also provide training samples for the system to automatically gather more instances for ratification. The system identifies opinion expressions as phrases containing opinion words, opinionated features and also opinion modifiers. These expressions are categorized as positive or negative with membership values varying from zero to one. Opinion expressions are identified and categorized using localized linguistic techniques. Opinions can be aggregated at any desired level of specificity i.e. feature level or product level, user level or site level, etc. We have developed a system based on this approach, which provides the user with a platform to analyze opinion expressions crawled from a set of pre-defined blogs.     

面向主题社团的意见领袖挖掘方法

意见领袖在不同的主题社团下对舆情的传播影响力是不同的,为了在社交网络中快速准确挖掘出意见领袖,提出一种面向主题社团的意见领袖挖掘方法。根据提出的兴趣隐含狄利克雷分布（Interest Latent Dirichlet Allocation,I-LDA）主题模型得到主题表达能力更强的主题分布,并在此基础上计算相邻用户的主题相似度。采用基于主题相似度的多标签均衡社团划分算法划分主题社团,使相似度大的用户被划分到相同的主题社团中,由此进一步提升社团划分的准确性与合理性。对于意见领袖的挖掘,提出一种快速意见领袖挖掘算法（Quickly-Ming Opinion Leader Algorithm,QMOLA）,先通过结构特征筛选出主题社团中的意见领袖候选人,再结合传播特征和情感特征挖掘主题社团中的意见领袖。对比实验结果表明,QMOLA相对于传统的意见领袖挖掘方法在挖掘效率上具有明显的优势,而且挖掘出的意见领袖具有更高的覆盖率和支持率。