CASIA OpenIR  > 毕业生  > 硕士学位论文
Alternative TitleComputational Modeling and Analysis of Netizen Groups: A Social Computing Approach
Thesis Advisor曾大军 ; 毛文吉
Degree Grantor中国科学院研究生院
Place of Conferral中国科学院自动化研究所
Degree Discipline模式识别与智能系统
Keyword网民群体 智能体建模 网络事件 行为抽取 情感分析 食品安全事件 Netizen Group Agent-based Modeling Internet Event Action extrAction Opinion Analysis Food Safety Event
Abstract网络事件是大量网民参与表达观点及评论的公共事件。作为一种新的网络社会现象,网络事件往往引发全国性关注并最终影响线下事件。参与网络事件的网民群体在这类事件中起着中心角色。网民群体的情感和观点,代表着网络社会舆情,是网络事件发展的关键因素。对网络事件中网民群体,尤其是其情感变化的社会建模,有助于了解网络事件的发展机制,为社会管理、决策支持提供有效参考。 本文聚焦于研究面向领域的网络事件中的网民群体及其情感演化,并提出了一个智能体模型来刻画网络事件中交互行为及其对网民群体状态的影响。实验基于两个中国网络事件进行案例分析。通过运行仿真来测试提出的模型并将实验结果与真实的社会媒体数据进行比较,展示了模型的有效性。 然而,传统的手工智能体建模方法费时费力。为了部分地减轻这一困难,本文以网络食品安全事件为例,提出利用计算方法从历史网络事件的文本信息中抽取领域知识来适应模型构建。首先,利用自然语言处理技术从历史网络事件报道中抽取各方行为构建智能体行为库,实验结果证实了方法的有效性;其次,提出了从网络事件评论中抽取网民情感并进行汇总以获得网民群体情感来支持建模的方法,该方法在实验评估中表现良好;最后,通过分析网络事件中的交互行为与网民群体情感状态变化的关联,修正了模型中的交互规则。 最后,本文利用计算方法抽取的领域知识重新构建了网民群体智能体模型,并将其与手工模型在真实网络食品安全事件上的模拟效果进行了比较。实验结果表明,通过计算方法构建的模型效果优于手工模型,可以用来完善手工模型。
Other AbstractInternet events are public events with the participation of netizens to express their opinions or comments. As an emerging Internet social phenomenon, Internet events often draw nationwide attention and eventually influence offline events. Netizen groups who participate in the Internet events play a central role in such events. Netizen group’s opinions and attitudes, which represent online public opinion, are the key factors to influence the evolution of Internet events. Social modeling of netizen groups, especially the dynamics of their opinions in Internet events, can help us understand the mechanism and evolvement of such events and provide valuable insights for social management and decision making. In this paper, we focus on the study of netizen groups and the dynamics of their opinions, and propose an agent-based model to capture the interactions in domain-oriented Internet events and their influence on the dynamics of netizen group’s states. Our experiment is based on two case studies of Chinese Internet events. We test the proposed model by running simulations and comparing experimental results with real social media data to show the effectiveness of our model. However, traditional hand-crafted agent-based model construction needs much human efforts and is time consuming. To partially alleviate this pain staking process, in this paper, we propose to employ computational techniques to extract domain knowledge from online textual data of historical events to facilitate the construction of agent-based model and illustrate the methods in food safety Internet events. First, we adopt multiple natural language processing techniques to extract the actions of every party from news reports of historical Internet events and construct the action library of the agents. The experimental results verify the effectiveness of the method. Second, we propose an approach to extract netizens’ opinions from comments of Internet events and summarize them to obtain netizen group’s opinions to support for model’s construction. It performs well in our experimental study. Last, we rectify the interaction rules in the model by analyzing the association between the interactions in Internet events and the dynamics of netizen group’s opinions. Finally, we reconstruct the agent-based model of netizen groups using the domain knowledge acquired by computational techniques, and compare its simulation results on real food safety Internet events with the hand-made model. Expe...
Other Identifier200928014628057
Document Type学位论文
Recommended Citation
GB/T 7714
谭章文. 网络事件中网民群体的社会计算建模与分析[D]. 中国科学院自动化研究所. 中国科学院研究生院,2012.
Files in This Item:
File Name/Size DocType Version Access License
CASIA_20092801462805(1609KB) 暂不开放CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[谭章文]'s Articles
Baidu academic
Similar articles in Baidu academic
[谭章文]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[谭章文]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.