复杂贸易网络结构分析与核心子网抽取算法研究

发布时间：2018-06-01 01:34

本文选题：复杂网络 + 带权网络　；参考：《东北大学》2014年硕士论文

【摘要】：现实世界当中,许多系统都是以复杂网络的形式存在的,世界贸易网络便是其中之一。世界贸易网络是一类时变的有权的网络,它是由世界各国之间贸易往来联系形成的复杂网络系统。分析其网络拓扑结构特征和网络核心子网及核心节点集对于预测世界贸易风险具有重要的意义。目前有权复杂网络的核心子网和核心节点集的计算方法存在诸多不足,首先是对衡量网络中节点重要性的公式设计依据不统一；其次是对核心子网和核心节点集的大小边界设定标准不唯一。因此本文以世界贸易网络为例进行分析,提出了一个新的求有权网络核心子网和核心节点集的算法。文中主要内容包含了两个部分：(1)通过实际的大规模国际贸易数据构建了多种不同类型的世界贸易网络,并详细的介绍了不同种类网络的构建方式。着重分析了各类贸易网络的拓扑结构特征并对贸易网络进行了社团分析,进而验证了贸易网络中的社团结构与当前的全球区域经济一体化相一致的结论。在此基础之上,给出了一个基于波达计数法算法求多层依赖网络中节点重要性排名的方法。(2)提出了通过依据连边权重阈值大小来删减网络连边,运用滑动窗口方法寻找网络达到稳定状态,从而得到网络的核心子网和核心节点集的算法。在此基础之上,文中也提出来一个依据统计方法网络核心节点集的方法。通过统计各个节点在网络的大量多层子网的核心子网中出现的频率,来标识节点重要性的方法。最后文中通过对多类贸易网络进行大量的实验,并通过与WTO的年度贸易报告给出的结果进行对比分析,验证了文中所提出的算法的有效性和正确性。
[Abstract]:In the real world, many systems exist in the form of complex networks, and the World Trade Network is one of them. The World Trade Network is a kind of time-varying and powerful network. It is a complex network system formed by the trade connection between countries. It is of great significance to analyze the characteristics of network topology and the core subnet and core node set for predicting the risk of world trade. At present, there are many shortcomings in the calculation methods of the core subnet and the core node set of the complex network. Firstly, the design basis of the formula to measure the importance of the nodes in the network is not uniform. Secondly, the standard of setting the size boundary of core subnet and core node set is not unique. Therefore, taking the World Trade Network as an example, this paper proposes a new algorithm for finding the core subnet and the core node set of the weighted network. The main content of this paper includes two parts: 1) based on the actual large-scale international trade data, this paper constructs a variety of different types of world trade networks, and introduces the construction of different types of networks in detail. The topological characteristics of various trade networks are analyzed emphatically and the association analysis of trade networks is carried out. The conclusion that the community structure in trade networks is consistent with the current global regional economic integration is verified. On the basis of this, a method of calculating the importance ranking of nodes in multi-layer dependent networks based on the algorithm of arrival counting is given.) the method of reducing the connected edges by the weight threshold of connected edges is proposed. The sliding window method is used to find the stable state of the network, and the algorithm of the core subnet and the core node set of the network is obtained. On this basis, a method of network core node set based on statistical method is also proposed in this paper. The method of identifying the importance of nodes is presented by counting the frequency of each node in the core subnet of a large number of multi-layer subnets of the network. In the end, a large number of experiments on multi-class trade networks are carried out, and the results are compared with the results given in WTO's annual trade report, and the validity and correctness of the proposed algorithm are verified.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：O157.5

【相似文献】