基于在线社交网络信息传播的重要用户发现

发布时间：2018-07-02 15:29

本文选题：权重Wap算法 + 节点重要性　；参考：《天津大学》2014年硕士论文

【摘要】：Web2.0网络时代的到来，带动了以新浪微博为代表的在线社交网络平台的迅速崛起。其用户数量随着市场规模的扩大不断激增。在线社交网络平台不仅仅是普通用户的交流平台。越来越多的企业在低成本和高利益的刺激下也纷纷加入。将对在线社交网络的研究与更为广泛的商业行为相结合，不仅能通过充分利用在线社交网络的低成本、海量用户、不受时间、空间、种族、文化等限制的优点为企业带来更高的经济效益，同时也能提高在线社交网络平台的用户体验。本文以新浪微博在线社交网络平台作为分析对象，通过对中国移动公司的一条官方微博进行分析，旨在获得该微博信息在其传播路径上的重要用户。本文以Wap序列分析算法为基础，提出并实现了一种改进的权重Wap算法。该权重Wap算法引入了节点权重参数，允许企业根据其广告营销成本，设定合理的节点权重阈值，，进而在构建权重Wap-Tree树型结构时有效过滤掉不符合节点权重阈值的节点用户。通过对节点权重参数的合理设置，能有效控制权重Wap算法挖掘出的频繁序列数量以及重要用户数量。与传统的图论角度的挖掘算法不同，本文没有假设信息传播路径最优，而是通过实验仿真的方式，生成了7万条尽可能贴近现实的信息传播路径序列数据，并提出了一种以树型结构来存储这些数据的思想，很好的展示了信息传播的方向性。文章通过语义分析算法对信息传播路径中的每个用户的评论进行分析，并将积极评论与消极评论的比例定义为节点权重，能够确保通过权重Wap算法挖掘出的重要用户对产品的宣传均是积极的、正面的。同时，文章提出了FileNet平台下企业申请数据挖掘和广告投放服务的自动化流程，实现了自动向重要用户发放广告的功能。经实验证明，权重Wap算法在现实意义、准确率与时间复杂度、信息传播方向性、剔除消极影响等方面均具有较大的优势。
[Abstract]:The advent of the Web 2.0 era has led to the rapid rise of online social networking platforms represented by Sina Weibo. The number of users along with the expansion of the market scale continues to surge. The online social network platform is not just the communication platform for ordinary users. More and more enterprises in low-cost and high-interest incentives have joined. Combining research on online social networks with broader business practices, not only by taking full advantage of the low cost of online social networks, a large number of users, regardless of time, space, race, The advantages of culture and other restrictions can bring higher economic benefits for enterprises, but also improve the online social network platform user experience. Based on Sina Weibo online social network platform, this paper analyzes an official Weibo of China Mobile in order to obtain the important users of the Weibo information in its transmission path. Based on the Wap sequence analysis algorithm, an improved weighted Wap algorithm is proposed and implemented in this paper. The weighted Wap algorithm introduces the node weight parameter, which allows the enterprise to set a reasonable weight threshold according to its advertising cost, and then effectively filter out the node users who do not conform to the node weight threshold when constructing the weight Wap-Tree tree structure. By setting the weight parameters reasonably, the number of frequent sequences and the number of important users can be effectively controlled by the weighted Wap algorithm. Different from the traditional mining algorithm of graph theory, this paper does not assume that the information transmission path is optimal, but generates 70,000 information transmission path sequence data which are as close to the reality as possible through experimental simulation. A tree structure is proposed to store these data, which shows the direction of information transmission. In this paper, the semantic analysis algorithm is used to analyze the comments of each user in the information transmission path, and the ratio of positive comments to negative comments is defined as the node weight. It can ensure that the important users' propaganda of the product is positive and positive by the weighted Wap algorithm. At the same time, the paper puts forward the automatic process of enterprise application data mining and advertising service under FileNet platform, and realizes the function of distributing advertisements to important users automatically. The experiments show that the weighted Wap algorithm has great advantages in practical significance, accuracy and time complexity, direction of information propagation, elimination of negative effects, and so on.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.092

【参考文献】