基于社交网络的话题传播模型挖掘及应用

发布时间：2018-03-16 06:51

本文选题：信息传播　切入点：爬虫系统　出处：《海南大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着智能手机、平台电脑等移动设备的快速发展,社交媒体呈现了爆发似的增长。用户可以在任何时间任何地点使用设备在社交媒体上发布自己的事件或者观点。社交媒体的广泛使用,使得数据量剧增。为了更好的确认社交媒体中信息的主题,社交媒体平台通过使用hashtag(主题标签)的方式标识信息的主题。这不仅方便了社交媒体对相同话题的信息聚类,也很大程度上帮助了用户在社交网络上更容易找到自己感兴趣的话题。研究话题在社交网络中的传播在很多方面都很有现实意义,例如市场营销等。针对这些情况,本文对社交网络中话题的传播模型进行了研究。主要包括以下几个方面的内容:首先根据需求分析,构建了面向微博的话题爬虫系统。以话题为中心,以广度优先遍历的算法爬取新浪微博中的话题微博数据。爬虫系统主要解决了动态网页数据解析的问题,以及礼貌爬取数据、人机验证等问题,有着稳定的爬取性能。然后对爬取的数据进行了统计分析。统计结果发现微博的只一次转发的数量占总转发数量的50%以上,说明话题微博数据有很高的聚集性。其次介绍了社交媒体中信息传播的模型—IC模型和TIC模型。IC模型和TIC模型假设信息以级联的方式在社交网络中传播。根据TIC模型的优势和微博数据的高度聚集性,本文提出了分层融合的话题传播模型—STIC模型。STIC模型将社交网络中的话题分文两层,第一层为信息只被转发一次的情况,第二层为信息被转发两次以上的情况。第一层数据以三大类特征作为输入特征值,使用SVM分类算法学习。第二层使用TIC模型学习。最后将两个模型的学习结果线性整合作为整个模型的学习结果。通过实验分析,STIC模型可以比TIC模型有更好的预测效果。
[Abstract]:With the rapid development of mobile devices such as smartphones, platforms and computers, Social media has grown like an explosion. Users can post their own events or opinions on social media at any time, anywhere, anywhere. In order to better identify the topic of information in social media, social media platform identifies the topic of information by using hashtag. this not only makes it convenient for social media to cluster information on the same topic, It also makes it easier for users to find topics of interest on social networks. Research on the spread of topics on social networks is of practical significance in many ways, such as marketing. This paper studies the communication model of topic in social network. It mainly includes the following aspects: firstly, according to the requirement analysis, a topic crawler system for Weibo is constructed. The crawler system mainly solves the problems of dynamic web page data parsing, politely crawling data, man-machine verification and so on. It has stable crawling performance. Then it makes a statistical analysis of the crawling data. The statistical results show that Weibo's number of forwarding only once accounted for more than 50% of the total number of retweets. It is shown that the topic Weibo data is highly aggregated. Secondly, the model IC model, TIC model .IC model and TIC model of information transmission in social media are introduced. The information is assumed to propagate in cascades in the social network. According to the TIC model, the paper introduces the model of information transmission in social media. Type A and Weibo's highly aggregated data, In this paper, we propose a hierarchical fusion topic propagation model-STIC model. STIC model divides the topic in social network into two layers. The first layer is the case that the information is forwarded only once. The second layer is the case where the information is forwarded more than twice. The first layer of data uses three categories of features as input eigenvalues. In the second layer, the learning results of the two models are linearly integrated as the learning results of the whole model. Through the experimental analysis, the prediction results of the SVM model can be better than that of the TIC model.
【学位授予单位】：海南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.1

【相似文献】