基于趋势分析的舆情监控系统

发布时间：2018-04-23 06:38

本文选题：舆情监测 + 蜘蛛技术　；参考：《电子科技大学》2013年硕士论文

【摘要】：随着互联网的快速发展,网络文化的安全问题也逐渐显现出来。网络舆情热点是网民思想情绪和群众利益诉求在网络上的集中反映,是网民热切关注的聚焦点,反映出一个时期网民的所思所想。各大网站的BBS作为互联网最活跃的平台之一,自然充斥着大量不良的、敏感的、反面的信息,如果能够建立对重点网站的自动监控平台,对敏感信息进行自动采集、归类、跟踪,筛选等,就能极大地减少资源投入,高效的开展舆情侦测工作。此外,对危害互联网的信息来源的挖掘与侦测也是我国目前网络安全工作的重点和难点。基于上述情况,本文首先对网络爬虫进行了初步探究,再把爬虫的原理和BBS特征结合在一起,提出了一个基于网络,BBS的舆情信息数据的采集和提取系统。本课题在对我国网络舆论尤其是群体性事件进行了深入分析研究的基础上,建立了基于趋势评估的舆情监控模型,设计并实现了基于趋势分析的舆情监控系统。具体进行了以下研究工作： 1、通过分析将整个舆情监控技术归纳为“采集技术”、“识别技术”和“判断技术”,并对各阶段的技术进行了研究。 2、对具有中国特色的网络舆论进行了分析和研究,总结了中国式的互联网舆论模式。 3、对如何判断舆论走向、如何识别负面趋势进行了研究。 4、构建了一种基于趋势评估的舆情监控模型：包括采集模块、浏览量识别模块、关键字识别模块、传播热度识别模块、聚类摘要模块、趋势分析模块,并进行了编码和实际应用。所建立的舆情监控系统实现了以下功能： 1、构建趋势评估模型：通过模糊数学分析对舆情这类模糊事物的性状进行可靠定量,计算出舆情发展趋势的具体值。 2、时间聚类过滤：对信息处理模块采用了时间聚类分析技术,过滤搜索引擎和数据采集监控点过时的信息,并依据监控点信息出现的时间变化与响应的增量值来加入趋势评估的建模。 3、通过趋势评估技术,将国内现有的舆情监控的采集模块和分析模块进行了关联。利用模糊综合评判的方法构建出舆情事件的可能性,结合贝叶斯网络技术评价舆情事件的发展态势。
[Abstract]:With the rapid development of the Internet, the security of the network culture has gradually emerged. The hot spot of network public opinion is the concentrated reflection of the netizens' thoughts and emotions and the interests of the masses on the network. It is the focal point that the netizens pay close attention to, and reflect the thoughts and thoughts of the netizens in a period of time. As one of the most active platforms on the Internet, the BBS of major websites is naturally filled with a large number of bad, sensitive and negative information. If we can establish an automatic monitoring platform for key websites, we can automatically collect and classify sensitive information. Tracking, screening, and so on, can greatly reduce the input of resources, efficient public opinion detection work. In addition, the mining and detection of the information sources that harm the Internet is also the focus and difficulty of the network security work in our country. Based on the above situation, this paper firstly probes into the web crawler, then combines the principle of crawler with the characteristics of BBS, and puts forward a collection and extraction system of public opinion information data based on BBS. On the basis of deeply analyzing and studying the network public opinion, especially the mass events in our country, this paper establishes a monitoring model of public opinion based on trend evaluation, and designs and implements a monitoring system of public opinion based on trend analysis. The following studies were carried out: 1. Through analysis, the whole monitoring technology of public opinion is divided into "collection technology", "identification technology" and "judgment technology", and the technology of each stage is studied. 2. This paper analyzes and studies the internet public opinion with Chinese characteristics, and sums up the Chinese internet public opinion mode. 3. How to judge the trend of public opinion and how to identify the negative trend are studied. 4. A monitoring model of public opinion based on trend assessment is constructed, which includes collection module, visitor recognition module, keyword identification module, propagation heat identification module, cluster summary module, trend analysis module, etc. Coding and practical application are carried out. The public opinion monitoring system realized the following functions: 1. Construct the trend evaluation model: through the fuzzy mathematics analysis to the public opinion this kind of fuzzy thing characteristic carries on the reliable quantification, calculates the public opinion development tendency concrete value. 2. Time clustering filtering: the information processing module adopts the technology of time clustering analysis to filter out the outdated information of search engine and data collection and control point. The model of trend assessment is added according to the increment value of time change and response of monitoring point information. 3. Through the trend evaluation technology, the collection module and the analysis module of public opinion monitoring are connected. The possibility of public opinion event is constructed by using fuzzy comprehensive evaluation method, and the development trend of public opinion event is evaluated with Bayesian network technology.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP393.08

【参考文献】