社交媒体中微博转发的预测模型研究

发布时间：2018-03-06 13:38

本文选题：微博转发　切入点：情感分析　出处：《北京邮电大学》2015年硕士论文　论文类型：学位论文

【摘要】：随着互联网和Web2.0技术的迅速发展和普及,互联网上出现了大量的在线社交网络。这些社交网络成为人们分享信息,传播信息,获取信息的主要平台。微博是一个基于用户关注粉丝关系的信息分享、传播以及获取平台。用户可以通过浏览器,智能手机客户端发布140个字以内的文字信息,图片和视频,实现即时分享。作为一个媒体平台,最重要的作用是信息的传播。在微博网络中,信息的传播主要是通过微博转发实现的。因此,微博的转发数量可以作为衡量微博传播效果的重要指标。研究信息在微博网络中的转发行为,对微博转发规模、数量的预测,在产品营销,热点提取和控制敏感信息的传播方面有重要作用和现实意义。本文从两个方面研究微博转发行为：1.微博能否被转发；2.微博的转发规模和数量。在微博能否被转发方面,本文从微博文本情感和用户角色两个方面入手,分析这两方面对微博转发行为的影响。在微博文本情感方面,本文构造了一个文本情感分析引擎来分析文本情感对微博转发行为的影响。在用户角色方面,通过对用户的聚类,验证了不同的用户角色的微博在转发行为方面的巨大差异。在研究微博的转发规模和数量方面,本文建立了两个微博转发量预测模型,分别是两阶段微博转发量预测模型和基于粉丝转发意愿和转发影响力的预测模型。在两阶段模型中,通过分类模型与回归模型的组合,有效的降低了数据不平衡的影响,得到了比较好的预测效果。通过分析微博转发量的产生机制,建立了基于粉丝转发意愿和粉丝转发影响力的微博转发预测模型,该模型有效率学习了微博转发量的增长过程,得到了比两阶段模型更好的预测效果。
[Abstract]:With the rapid development and popularization of the Internet and Web2.0 technology, a large number of online social networks have emerged on the Internet. Weibo is a platform for sharing, disseminating and acquiring information based on the user's focus on fan relationships. Users can publish 140 characters of text information, pictures and videos through browsers and smartphone clients. Realize instant sharing. As a media platform, the most important role is the dissemination of information. In Weibo's network, the dissemination of information is mainly realized through the transmission of Weibo. Therefore, Weibo's quantity of retweets can be used as an important indicator to measure the effect of Weibo's dissemination. Research on the behavior of forwarding information in the Weibo network, the prediction of the scale and quantity of the forwarding of Weibo, and the marketing of the products. Hot spot extraction and control of the dissemination of sensitive information play an important role and practical significance. This paper studies the forwarding behavior of Weibo:: 1. Whether Weibo can be forwarded. 2. The scale and quantity of the retweeting of Weibo. In terms of whether Weibo can be forwarded, this paper analyzes the influence of the two aspects on Weibo's retweeting behavior from two aspects: the emotion and the role of the user. In this paper, a text emotion analysis engine is constructed to analyze the influence of text emotion on Weibo's forwarding behavior. In the aspect of studying Weibo's forwarding scale and quantity, this paper sets up two forecasting models of Weibo's forwarding quantity. The two stage Weibo forwarder prediction model and the prediction model based on fan forwarding will and forwarding influence respectively. In the two-stage model, through the combination of classification model and regression model, the effect of data imbalance is effectively reduced. By analyzing the generation mechanism of Weibo's retweeting quantity, we set up a Weibo forwarding prediction model based on fans' forwarding will and fan's forwarding influence. This model can effectively learn the process of the growth of Weibo's retweeting quantity. The prediction results are better than that of the two-stage model.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP393.092

【参考文献】