基于阈值的金融网络的构建及其应用研究

发布时间：2018-06-06 11:35

本文选题：金融网络 + 有效阈值　；参考：《武汉理工大学》2013年硕士论文

【摘要】：金融市场是一个高度复杂的动力学系统,基于不同股票之间的相关性构建的金融网络有助于理解金融系统的内在结构和动力学行为。基于复杂网络研究并分析金融市场的结构及其性质需要建立在构建网络的基础上。构建网络的方法有多种,其中边相关系数阈值法被广泛应用,但目前学者们在基于边相关系数阈值法构建网络时大多是考虑采用传统技术即选取固定的阈值的方法,然而研究发现,选择不同的阈值得到的网络的拓扑结构有着明显的差异。本文首先论证了最大连通子图在金融网络中的地位,然后提出了有效阈值区间的概念,最后给出了基于阈值构建金融网络新方法的具体步骤,并将该方法应用到实际中去。本文基于阈值研究金融网络的构建及其应用研究,主要表现在以下几个方面。 (1)通过分析不同阈值下最大连通子图的节点个数的变化情况,发现最大连通子图是金融网络的共性,是金融网络的内在属性。 (2)通过分析股票的皮尔逊相关系数分布情况,发现均呈现一定的正态性,由此确定金融网络在构建时的有效阈值区间,并根据最大连通子图的节点个数的变化情况进一步来确定构建网络的阈值。 (3)给出基于阈值构建金融网络的新方法。最后,将该方法应用到实际中去,分别构建了沪市金融网络和道琼斯金融网络。分析了沪市金融网络的模体结构和社团结构,并基于模体分析了金融市场中投资者购买股票的可能性；然后,分析了不同阈值下道琼斯金融网络的稳健性变化和投资风险变化情况,进一步探讨了阈值选取的问题。
[Abstract]:The financial market is a highly complex dynamic system. The financial network based on the correlation between different stocks is helpful to understand the internal structure and dynamic behavior of the financial system. To study and analyze the structure and nature of financial market based on complex network, we need to build a network. There are many methods to construct the network, among which the threshold method of edge correlation coefficient is widely used. But at present, most scholars consider the method of choosing fixed threshold when constructing the network based on the threshold method of edge correlation coefficient. However, it is found that there are significant differences in the topology of the network with different thresholds. In this paper, the position of maximum connected subgraph in financial network is first demonstrated, then the concept of effective threshold interval is proposed. Finally, the concrete steps of constructing a new method of financial network based on threshold are given, and the method is applied to practice. This paper studies the construction and application of financial network based on threshold, mainly in the following aspects. 1) by analyzing the change of the number of nodes in the maximum connected subgraph under different thresholds, it is found that the maximum connected subgraph is the common character of the financial network and the intrinsic attribute of the financial network. 2) by analyzing the Pearson correlation coefficient distribution of stocks, it is found that all of them are normal, and the effective threshold range of financial network is determined when the financial network is constructed. The threshold of constructing network is further determined according to the change of the number of nodes in the maximum connected subgraph. A new method of constructing financial network based on threshold is presented. Finally, the financial network of Shanghai Stock Exchange and the financial network of Dow Jones are constructed by applying this method to practice. This paper analyzes the motif structure and community structure of the financial network of Shanghai Stock Exchange, and analyzes the possibility of investors buying stocks in the financial market based on the motifs. This paper analyzes the changes of robustness and investment risk of Dow Jones financial network under different thresholds, and further discusses the problem of threshold selection.
【学位授予单位】：武汉理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F830.91;O157.5

【参考文献】