基于微博公共平台的单条微博热门程度预测及其相关微博发现

发布时间：2018-06-11 13:37

本文选题：微博 + 热门话题　；参考：《安徽大学》2016年硕士论文

【摘要】：新浪微博在我们日常生活的信息传播中起到了重要的作用,截止2014年新浪微博的每日活跃人数已经达到了6660万。而新浪微博公共平台一般拥有过万的粉丝数,他们的影响力及其大,因此对于新闻的监管就变得十分的困难,如何实时从成千上万的微博中找到热门的微博是十分有意义的。它可以帮助人们实时的了解社会动态,使一些政府机构提前对一些负面的情绪加以引导,从而使得社会更加的稳定。在当前微博发展的趋势下,学者们对微博的研究也越来越重视,最近几年对微博的热门话题发现、微博的舆情分析都是研究的热门。本论文主要构建了一个微博公共平台实时的监管系统,用来预测将来可能会成为热门话题的微博。本论文主要对新浪微博的各大新闻公共平台进行数据采集,然后对这些数据进行分析,通过这样的分析本文构建了预测模型。通过计算每条微博的转发、评论的增长趋势,从而预测出该条微博在当前的热门程度。并且通过聚类算法找出与该微博相关的一些微博。本论文的主要贡献如下：(1) 通过新浪微博api接口实时收集微博数据,本文通过使用新浪微博api接口每十分钟收集一次新浪微博的数据,然后对这些微博进行分析,找到影响微博趋势的因素,通过这样的分析发现微博的评论、转发增长数会随着时间的推移呈现凹曲线下降的趋势。(2) 本论文构建了一个预测模型,该模型由微博的长期趋势、循环趋势、作者影响力、微博前十分钟的评论数和转发数几个因素组成。通过该模型可以计算出一条微博的评论数和转发数在未来的趋势,然后本文给出了热门程度公式,通过这个公式可以计算一条微博的热门程度。(3) 本论文改进了k-means聚类算法,该算法通过输入一个聚类中心并且设置一个阈值来找到距离聚类中心在这个阂值内的文本。通过这样的算法可以找到与热门微博相关的一些微博。
[Abstract]:Sina Weibo has played an important role in the dissemination of information about our daily lives, with the daily active number of Sina Weibo reaching 66.6 million as of 2014. Sina's Weibo public platform generally has more than 10, 000 followers, and their influence is enormous, making it difficult to monitor news, and it makes sense to find popular Weibo from thousands of Weibo in real time. It can help people to understand the social dynamics in real time, so that some government agencies in advance to guide some of the negative emotions, thereby making the society more stable. With the development of Weibo, scholars pay more and more attention to the research of Weibo. In recent years, the popular topic of Weibo is that the analysis of public opinion of Weibo is the hot research. This paper mainly constructs a real-time monitoring system of Weibo public platform to predict the Weibo which may become a hot topic in the future. This paper mainly carries on the data collection to the Sina Weibo each big news public platform, then carries on the analysis to these data, through this kind of analysis this article has constructed the forecast model. By calculating the forward of each Weibo and the growth trend of the comments, the current popularity of the Weibo is predicted. Some Weibo related to the Weibo are found by clustering algorithm. The main contributions of this paper are as follows: (1) collect Weibo data in real time through Sina Weibo api interface. This paper collects Sina Weibo data every ten minutes by using Sina Weibo api interface, and then analyzes these Weibo. Find out the factors that influence the trend of Weibo, through this analysis we find that the comments of Weibo, the number of forward growth will show a downward trend of concave curve over time.) this paper constructs a prediction model, which is based on the long-term trend of Weibo. Circular trends, author influence, Weibo's first 10 minutes of comments and forwarding a number of factors. Through this model, we can calculate the future trend of the number of comments and forwards of a Weibo. Then, this paper gives the popular degree formula, which can be used to calculate the popularity of a Weibo.) in this paper, we improve the k-means clustering algorithm. By inputting a cluster center and setting a threshold value, the algorithm finds the text within the threshold value of the distance clustering center. Through this algorithm, we can find some Weibo related to the popular Weibo.
【学位授予单位】：安徽大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP393.092

【相似文献】