微博热点话题传播模型与可视化研究

项目名称： 微博热点话题传播模型与可视化研究

项目编号： No.61272367

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 叶施仁

作者单位： 常州大学

项目金额： 80万元

中文摘要： 微博作为一种新型的即时通信与信息共享服务，造成的"发布革命"让中文微博用户数已经达超过3亿。怎样预见性地发现和检索那些具有广泛传播能力的热点话题，以便尽可能早地掌握情况，具有重大的社会意义和经济效益。本项目旨在全面探索中文微博的基本特征的基础上，研究微博信息快速传播的机制和适合微博特征搜索模型。我们对新浪微博为代表的中文微博的网络结构进行全面分析，为学术界和产业界提供微博数据、微博处理工具和基本结论等系列资源。我们通过情感分析研究微博中的话题性，探索主题传播速度和传播范围的机制并建立能处理海量数据的变粒度传播模型。我们研究反映传播趋势的微博搜索模型，搜索的排名通过机器学习的方法确定确定博主、内容和博主之间交互的不同贡献。搜索返回的相关发帖列表综合考虑了它们的话题性、权威性与新颖性。我们研究可视化的模型来跟踪相关话题的发展过程和搜索的结果，为用户提供快速理解微博数据传播的途径。

中文关键词： 社会媒体；微博；社会网络；水军；情感分析

英文摘要： Microblogs have been leading a publishing revolution as novel instant messaging and information-sharing services. Nowadays there are over 300 millions microblog users in China. If we can detect and retrieve the relevant blogs that tend to propagate heavily soon, we will understand the whole hot topic as early as possible. The predictable search results will be beneficial to industrial applications and social researches. The project will explore the fundamental characteristics, investigate propagation mechanism and microblog-customised retrieval model of Chinese microblogs. We use Sina Weibo, a typical and popular Chinese microblog site, to reveal the network structure features of Chinese microblog, and provide test corpora, processing tools and baselines for research and industry community. We build propagation model based on sentiment analysis to evaluate the topicality, as well as propagation speed and scope. The propagation model considers bloggers with various granularity in order to handle huge bloggers existing in Sina Weibo. We also develop the microblog search model indicating propagation tendency, where the search ranking is learned from the features tied to authors and their blogs, as well as interactions between authors. The returned blogs will be ranked according to their topicality, authority and no

英文关键词： Social Media；Microblogging；Social Network；Water Army；Sentiment Analysis