跨语言社会舆情分析基础理论与关键技术研究

项目名称： 跨语言社会舆情分析基础理论与关键技术研究

项目编号： No.61331013

项目类型： 重点项目

立项/批准年度： 2014

项目学科： 无线电电子学、电信技术

项目作者： 赵小兵

作者单位： 中央民族大学

项目金额： 285万元

中文摘要： 随着网络技术的发展，蒙藏维文网站增长迅猛、内容丰富。各民族网民实时通过论坛、博客等发表言论，形成网络舆论。只有及时掌握舆情动向，才能正面引导舆论和宣传。本课题通过动态监测蒙藏维汉民族语言网络资源，建立涵盖宗教、经济、科教等领域的社会舆情综合语料库；应用跨语言社会舆情监测指标体系，建立统一的集成管理平台，对少数民族地区关注的热点或敏感话题进行识别与动态跟踪，通过话题的倾向性分析，实现少数民族语言网络资源的群体行为分析及综合舆情监测。本研究重点解决的关键技术问题包括：多民族文字共现的网络数据实时下载、编码转换、词法分析等预处理技术；面向舆情监测的多民族语言语义本体知识库建设；基于语义本体的跨语言信息检索；基于跨语言时间演化主题模型的热点话题检测与跟踪、文本倾向性分析等技术，形成民族地区舆情分析报告，为决策层全面掌握民族地区舆情动态，做出正确舆论引导，实现提供技术支撑和接近实际使用需求的系统。

中文关键词： 舆情分析;智能信息处理;少数民族语言;跨语言;

英文摘要： The number of Mongolian, Tibetan and Uyghur websites is growing rapidly with more and more details. Every user can declare his issues through BBS, blogs and other network medias from time to time, forming network public opinion. To positively guide public opinions and propaganda, we need to find the hot topics timely.. In this project, we will establish public opinion corpus in religion, economy, science and other fields by collecting information through monitoring minority language network resources dynamically. We will also construct an integrated management platform, in terms of the cross language social public opinion monitoring standard, to recognize and track the hot and sensitive topics in national regions. We will then carry out group behavior analysis and public opinion monitoring using text orientation analysis. . The key technologies in our research include:. (1) Minority language text preprocessing technology, such as network real-time data download, code conversion, morphology analysis, etc.. (2) Minority language semantic ontology knowledge base construction for public opinion monitoring.. (3) Public opinion monitoring analysis technique based on the semantic ontology, such as cross language information retrieval, hot topic identification and tracking based on continuous time dynamic topic models, text orientation analysis etc.. With all of these technologies, the minority area public opinion analysis report will be formed for government to grasp public opinion and make positive guidance. The research output of this project can provide methodological and technical support for realizing above functions.

英文关键词： Analysis of Public Opinion;Intelligent Info Processing;Minority Language;Cross Language