一种基于文本语义的网络敏感话题识别方法

网络已成为当今世界重要的信息载体,但是网络信息良莠不齐,对人们的生活造成了很多负面影响,因此,如何正确识别网络中的敏感话题,是当前网络舆情分析与监管的重要任务之一。本文以识别网络论坛中的敏感话题为目标,基于网络论坛文本在结构和表达上表现出的篇幅短、结构不完整、文字口语化等特性,将该类文本表示成基于向量空间模型的文本矩阵,并根据网络敏感话题具有先验知识和态度倾向性等特点,提出了基于倾向性词典的特征提取方法,可有效提高网络敏感话题识别的正确率,最后通过实验验证了这一改进的有效性,证实了本文的研究价值。     

科技文献话题演化研究

提出一种研究话题演化的方法,利用LDA话题模型抽取科技文献的话题,通过计算话题的强度和特征词,研究话题的演化趋势。对NIPS论文集与ACL论文集进行实验,结果显示了机器学习领域以及计算语言学领域的一些发展状况,从而验证该方法的可行性。     

Strategic topic avoidance: an investigation of topic avoidance frequency,strategies used,and relational correlates

This research stresses the need to examine the relationship between topic avoidance and relational correlates (e.g., satisfaction and emotional closeness) from a message production theoretical perspective. Our approach—strategic topic avoidance—offers additional explanatory capabilities as the strategies with which interactants in close relationships avoid topics may be associated with perceptions of the relationship (after accounting for topic avoidance frequency). Moreover, relational correlates may also vary by the combination of overall topic avoidance frequency and certain topic avoidance strategies. The current research, therefore, assessed individuals' topic avoidance frequency levels and the frequency of using topic avoidance strategies in relation to satisfaction and closeness across three different relational types (i.e., significant others, mother–young‐adult, and father–young‐adult relationships). Results suggested that avoiding certain topics, such as current relational concerns, predicted levels of satisfaction and closeness across relationship types; however, cross‐relational differences also emerged. Strategies employed to avoid topics accounted for additional variance in satisfaction and closeness for relationships with significant others and mothers but not fathers. Analyses also demonstrated that overall topic avoidance frequency interacted with topic avoidance strategy use.     

重大科技基础设施对学科领域发展影响的计量评价研究——以对撞机为例

重大科技基础设施（简称重大设施）是具有前沿性、创新性、规模性及影响力的科学研究系统,在国家科技创新体系中占据重要地位。为了深刻认识重大设施的战略价值、充分发挥设施科研潜力、支撑未来设施的统筹布局,需要深入研究其对学科领域发展的影响作用。以专用型重大设施中的对撞机为研究对象,分析其论文成果的引文分布及主题强度演化,研究了重大设施对学科领域发展的影响过程,定义了不同影响类型,并形成了判别不同类型的三维战略坐标。在对比中、日、欧三方对撞机的成果产出和影响情况后,提出了相关政策建议。     

Exploring the effect of task instructions on topic beliefs and topic belief justifications: A mixed methods study

The purpose of this mixed methods study was to investigate whether task instructions that asked adolescents to evaluate the merit of both sides of a controversial issue would affect their topic beliefs and topic belief justifications after they read belief-consistent and belief-inconsistent information. In the quantitative phase, we conducted an experiment in which high school students (n = 45) were randomly assigned to one of four conditions and received their respective pre-reading task instructions. Quantitative analyses showed that task instructions affected topic beliefs and belief justifications. However, inspection of topic belief scores within each condition indicated that some individuals’ beliefs became weaker, whereas others’ became stronger. In the qualitative phase, we conducted interviews to explain why this occurred. The interview data revealed two distinct reader profiles: belief-reflection and belief-protection. The data sets were complementary: the quantitative data indicated group differences in topic beliefs and belief justifications, and the qualitative data allowed us to explain differences within and across groups.     

"I'd rather not talk about it": adolescents' and young adults' use of topic avoidance in stepfamilies

This study examined adolescents' and young adults' use of topic avoidance with their mothers, fathers, stepmothers, and stepfathers. The types of topics avoided differed according to the type of parent-child relationship. Specifically, adolescents and young adults engaged in the most topic avoidance with their stepparents (regardless of whether the stepparent was a stepmother or stepfather), followed by their fathers, and then their mothers. Quantitative measures indicated that sex was the most frequently avoided topic across all relationship types. Open-ended responses revealed additional commonly avoided topics, including talking about the other parent/family, deep conversations, and money (e.g., child support payments). The most frequently reported reasons for this avoidance were self protection, relationship protection, and conflict. This research suggests that children in stepfamilies face unique decisions about topic avoidance. Communication Boundary Management Theory (Petronio, 1991) was used to explain how adolescents and young adults might engage in topic avoidance to regulate their personal boundaries, constructing relatively impermeable boundaries with some adults while maintaining looser boundaries with others. Finally, numerous practical suggestions are offered for understanding the balance between openness and closedness in stepfamilies and for promoting healthy stepfamily functioning.     

Topic Avoidance,Everyday Talk,and Stress in Romantic Military and Non-Military Couples

Relationships are increasingly impacted by military involvement, necessitating a deeper understanding of communication within these couples. This study examined perceptions of everyday talk, topic avoidance, and stress of participants romantically involved with (a) a military partner (n = 118) or (b) a non-military partner (n = 94). Couple types avoided similar topics, but non-military partners avoided these topics more frequently than military partners. Both couple types engaged in everyday talk, but military partners found this communicative exchange to be more important. Non-military partners experienced greater stress than military partners, and communication was predictive of the stress experienced. The results of this study highlight similarities and differences in romantic couples and reveal areas in which relationship education can address these communicative nuances of military couples.     

Topic Valence and Ambiguity in Parent-Emerging Adult Child Postdivorce Discussions

Communication privacy management theory maintains permeability rules manifest as communication strategies used during conversation. Although postdivorce families tend to privilege openness, this conclusion is based on recalled reports and single discussions about negatively valenced topics. To determine whether ambiguity functions as a manifestation of permeability rules and its effects on relational satisfaction and psychological well-being, 39 parent-emerging adult child dyads discussed one negatively valenced and one positively valenced postdivorce topic. Results indicated parents were ambiguous while discussing negatively and positively valenced topics, and children were perceptive of parents' ambiguity during the conversations. No differences were found between parents' use and children's perceptions of parents' ambiguity. Further, children's relational satisfaction and psychological well-being were affected by ambiguity in both discussions.     

面向教育技术学文献数据的主题挖掘

对网络环境下海量的科技文献数据进行文本挖掘可以有效地提高科技文献信息的可用性,发现隐藏的知识。LDA (Latent Dirichlet Allocation)模型是一种能够提取文本隐含主题的非监督学习模型。论文基于LDA模型,以三种国际教育技术期刊在2004-2008年间出版的学术文献为研究对象,进行了主题挖掘和文献分析。     

Content analysis of e-petitions with topic modeling: How to train and evaluate LDA models?

E-petitions have become a popular vehicle for political activism, but studying them has been difficult because efficient methods for analyzing their content are currently lacking. Researchers have used topic modeling for content analysis, but current practices carry some serious limitations. While modeling may be more efficient than manually reading each petition, it generally relies on unsupervised machine learning and so requires a dependable training and validation process. And so this paper describes a framework to train and validate Latent Dirichlet Allocation (LDA), the simplest and most popular topic modeling algorithm, using e-petition data. With rigorous training and evaluation, 87% of LDA-generated topics made sense to human judges. Topics also aligned well with results from an independent content analysis by the Pew Research Center, and were strongly associated with corresponding social events. Computer-assisted content analysts can benefit from our guidelines to supervise every process of training and evaluation of LDA. Software developers can benefit from learning the demands of social scientists when using LDA for content analysis. These findings have significant implications for developing LDA tools and assuring validity and interpretability of LDA content analysis. In addition, LDA topics can have some advantages over subjects extracted by manual content analysis by reflecting multiple themes expressed in texts, by extracting new themes that are not highlighted by human coders, and by being less prone to human bias.