基于中文微博的突发话题检测系统的设计与实现

发布时间：2018-06-02 19:30

本文选题：中文微博 + 突发话题检测　；参考：《北京邮电大学》2014年硕士论文

【摘要】：随着社会化网络服务(Social Network Services, SNS)的不断发展,微博(Microblogging)已经成为了很多国人生活中一个必不可少的组成部分。微博是一种通过单向的关注机制来分享简短即时消息的广播式的社交网络平台。作为一种社会化媒体,它所具有的短文本、实时、社交和媒体特性大大缩短了信息从发布到广泛扩散蔓延的时间,更加有利于信息的快速传播。当前,微博已经成为了网络舆论的主要爆发地和聚集地。有效地检测出微博中的突发话题,不管是对普通用户、商家还是政府部门来说,都有着很强的现实意义。本论文在总结现有突发话题检测研究成果的基础上,设计并实现了一个基于新浪微博的突发话题检测系统(Emerging Topic Detection System),简称ETD。它实时地从新浪微博采集用户和微博数据,并尽可能提高数据集的完整性与一致性。它使用了一种新颖的的突发话题检测模型,能够更准确地将大量微博中的突发话题检测出来。最后通过使用一些最新的数据可视化技术,将检测出的突发话题信息在Web前端进行展示。论文首先对突发话题检测的相关理论和技术背景进行了简要介绍；之后详细描述了基于中文微博的突发话题检测系统ETD的需求分析,并对整个系统进行了总体设计；接下来描述了系统内部各个子系统的详细设计与实现,包括数据采集子系统、突发话题检测子系统和突发话题可视化子系统；然后,对整个系统进行单元测试和集成测试,表明整个系统达到了预期的设计目标；论文最后对全文进行了总结,对未来的工作进行了展望,并总结了作者在研究生期间的所有工作和成果。
[Abstract]:With the development of social Network Services, SNS), Weibo microblogging has become an essential part of Chinese life. Weibo is a broadcast social networking platform that shares short instant messages through a one-way focus mechanism. As a kind of social media, it has the features of short text, real-time, social and media, which greatly shorten the time of information spreading from publication to wide spread, and is more conducive to the rapid dissemination of information. At present, Weibo has become the main outbreak of network public opinion and gathering. It is of great practical significance to detect the burst topic in Weibo effectively, whether for ordinary users, merchants or government departments. On the basis of summarizing the existing research results of burst topic detection, this paper designs and implements an emerging Topic Detection system based on Sina Weibo. It collects user and Weibo data from Sina Weibo in real time, and improves the integrity and consistency of data set as much as possible. It uses a novel burst topic detection model, which can detect a large number of burst topics in Weibo more accurately. Finally, by using some latest data visualization techniques, the detected burst topic information is displayed in the front end of Web. Firstly, this paper briefly introduces the theory and technology background of burst topic detection, then describes the requirement analysis of burst topic detection system (ETD) based on Chinese Weibo in detail, and designs the whole system. Then it describes the detailed design and implementation of each subsystem of the system, including data acquisition subsystem, burst topic detection subsystem and burst topic visualization subsystem. It shows that the whole system has achieved the expected design goal. Finally, the thesis summarizes the full text, prospects the future work, and summarizes all the work and achievements of the author during the postgraduate period.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.092

【参考文献】