瞬态社会网络中信息扩散与影响力最大化

发布时间：2018-08-25 09:40

【摘要】：瞬态社会网络是指在特定时间、为特定事件、持续时间短及面对面接触所形成的社会网络。不同于在线社会网络,瞬态社会网络由于其特征,能够提供更加安全可靠的信息,但是瞬态社会网络也有自己的缺点,持续时间太短,稍纵即逝,对于这样寿命很短却有效的社会网络,在其中做信息扩散的研究是一件很有意义的事情,但是,对于这种社会网络中信息扩散的研究,传统的方法直接用在其中已经是不可能的事情了,因此,怎样依据在线社会网络中的信息扩散的研究成果来研究瞬态社会网络中的信息扩散也就成了研究的方向。本文首先对瞬态社会网络特征及研究现状进行了分析,指出了已有的研究成果并不能完全考虑社会网络的特征。然后,对信息扩散及信息扩散最大化的相关进展进行了较为详实的介绍与分析。针对瞬态社会网络中信息扩散的问题,提出了一个基于结构洞的信息扩散模型。首先,现有的方法通常是针对在线社会网络的,并不能直接运用到瞬态社会网络中。其次,结合在线社会网络中的信息扩散模型,是否有能够满足瞬态社会网络特征的模型。最后,以真实数据集为基础,进行实验验证模型的正确型与有效性。瞬态社会网络的易变性导致自身的存在时间非常短,如果就某一时刻的瞬态社会网络进行研究太过狭窄,我们就结合结构洞,整合了所有时刻的瞬态社会网络组成全局瞬态社会网络,这样的话,我们就可以像沿用在线社会网络的方法进行研究。因此,瞬态网络与其他的任何网络一样,只要能够找到网络的特征,就能够改进我们的模型以适用瞬态社会网络。随后,我们继续进行信息扩散最大化的研究,针对瞬态网络中的相遇时间和相遇次数特征来进行分析。首先,提出了瞬态社会网络中的相关概念并给出了问题的定义。然后,设计基于相遇次数和相遇时间来确定节点的影响力,在全局瞬态社会网络下,找到影响力最大的节点集合并作为初始受众,有效的解决了瞬态社会网络中信息扩散最大化问题。最后,在两个真实数据集上进行了实验,验证所提出算法的可行性及有效性。
[Abstract]:Transient social network is a kind of social network formed by specific events, short duration and face-to-face contact at a particular time. Unlike online social networks, transient social networks can provide more secure and reliable information because of their characteristics, but transient social networks also have their own shortcomings, too short duration, fleeting, It is very meaningful to study the diffusion of information in such a short but effective social network, but for the study of the diffusion of information in such a social network, It is impossible for traditional methods to be directly used among them. Therefore, how to study the diffusion of information in transient social networks based on the research results of information diffusion in online social networks has become the research direction. In this paper, the characteristics and research status of transient social networks are analyzed, and it is pointed out that the existing research results can not fully consider the characteristics of social networks. Then, the development of information diffusion and information diffusion maximization is introduced and analyzed in detail. In order to solve the problem of information diffusion in transient social networks, a structural hole based information diffusion model is proposed. Firstly, the existing methods are usually aimed at online social networks and can not be directly applied to transient social networks. Secondly, combining the information diffusion model in the online social network, whether there is a model that can satisfy the characteristics of the transient social network. Finally, based on the real data set, the model is verified by experiments. The variability of transient social networks leads to a very short time of existence. If it is too narrow to study transient social networks at any given time, we will combine the structural holes. By integrating transient social networks at all times to form global transient social networks, we can study them like online social networks. Therefore, as with any other network, as long as we can find the characteristics of the network, we can improve our model to adapt to the transient social network. Then we continue to study the maximization of information diffusion and analyze the characteristics of encounter time and encounter times in transient networks. Firstly, the concept of transient social network is proposed and the definition of the problem is given. Then, the influence of nodes is determined based on the number of encounters and the time of encounters. In the global transient social network, the set of the most influential nodes is found and used as the initial audience. It effectively solves the problem of information diffusion maximization in transient social networks. Finally, experiments are carried out on two real data sets to verify the feasibility and effectiveness of the proposed algorithm.
【学位授予单位】：江西财经大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.09

【参考文献】