一种基于拓扑势的社会网络节点影响力评估方法

发布时间：2018-06-02 01:43

本文选题：社会网络 + 节点重要性　；参考：《哈尔滨工程大学》2016年硕士论文

【摘要】：伴随着社会经济和互联网技术的飞速发展,人们的生活方式和交流手段出现了革命式的变革,特别是在移动互联网技术快速进步、智能终端大量普及的背景下,人类社会形态表现出网络化的趋势,人与人之间的沟通也变得越来越频繁和紧密,信息传播的速度越来越快,逐步形成了以人与关系为交织的社会网络,并呈现出复杂、动态、规模大和自相似的特性。针对社会网络的研究与分析已经广泛应用到信息、经济、军事以及安全等领域,由于社会网络的跨学科、复杂以及动态等特性,对其进行深入分析和掌握存在着一定的难度和挑战,特别是分别从微观和宏观上完成节点重要性的评估和社区结构的识别,具有非常重大的实际意义和应用价值。针对以往对社会网络的分析大多数都是在简单图上进行的,在节点影响力评估中没有考虑节点和节点间关系的固有属性以及无法表达多方关系的问题,利用超图理论构建社会网络的超图模型,将节点和它们之间联系的固有属性引入到模型当中。然后引入认知物理学中数据场理论中的节点拓扑势来量化节点的重要性来表征影响力,分别利用熵权法、弗洛伊德算法进行节点质量和最短路径的计算,结合对影响因子的理想取值的分析,提出一种社会网络超图模型下改进的基于删除策略的节点重要性评估算法。最后,将基于节点重要性的评估方法应用于社区识别中,采用异步思想构建基于节点重要性增益的随机游走概率模型和可达策略,引入基于随机游走的节点相似性度量函数和迭代更新函数,提出一种基于社会网络节点重要性增益的层次社区识别算法作为节点重要性排序的应用和评估算法的性能的验证,并通过实验来验证算法的正确性和可行性。
[Abstract]:With the rapid development of social economy and Internet technology, people's way of life and means of communication have undergone revolutionary changes, especially in the context of the rapid progress of mobile Internet technology and the widespread use of intelligent terminals. The form of human society shows the trend of networking, the communication between people becomes more and more frequent and close, the speed of information dissemination is getting faster and faster, and gradually formed the social network which is intertwined with people and relations, and presents complex and dynamic. Large-scale and self-similar characteristics. The research and analysis of social network has been widely used in the fields of information, economy, military and security. Due to the interdisciplinary, complex and dynamic characteristics of social network, There are some difficulties and challenges in analyzing and mastering them, especially the evaluation of node importance and the identification of community structure from the micro and macro aspects, which have great practical significance and application value. In view of the fact that most of the analysis of social network in the past is based on simple graph, the inherent attributes of the relationship between nodes and the problem of multi-party relations can not be considered in the evaluation of node impact. The hypergraph model of social network is constructed by using hypergraph theory, and the inherent attribute of the connection between nodes and them is introduced into the model. Then the node topological potential in data field theory in cognitive physics is introduced to quantify the importance of nodes to represent the influence. The entropy weight method and Freud algorithm are used to calculate the quality of nodes and the shortest path, respectively. Based on the analysis of the ideal values of the influence factors, an improved node importance evaluation algorithm based on deletion strategy under the hypergraph model of social networks is proposed. Finally, the evaluation method based on node importance is applied to community identification, and the random walk probability model and reachable strategy based on node importance gain are constructed by asynchronous method. A hierarchical community recognition algorithm based on node importance gain of social network is proposed by introducing random walk based node similarity measure function and iterative update function as the application of node importance ranking and the performance verification of the algorithm. The correctness and feasibility of the algorithm are verified by experiments.
【学位授予单位】：哈尔滨工程大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：O157.5;TP301.6

【相似文献】