Research On Distinguishing The Clusters Of Capital Abnormal Flow In Financial Network

The financial network is organic synthesis of the financial activity, information and network technology. In the finance cyberspace, the special paths or flows may be formed between the two certain vertexes which can reflect the different economical relationship. The criminal activities, such as money laundering and the capital escaping, as well as hot money exceptionally flows cannot be implemented without the financial network.In the financial network, all kinds of unusual fund flowing threaten the credit of national financial organization: draining the national wealth massively and increasing the social unstable factors. Therefore, we must strengthen supervising and monitoring each kind of unusual funds to guarantee the stability of national economy and social order.All forms of complex systems in the nature can be described by network. The research of complex network theory already has become the most advanced and the hot spot now. The studies on the problems of complex financial network can be based on theories of complex network. It is a new research domain that studies the financial network by comparing the general character of network and the individuality of financial network.The introduction of this paper points out the goal of the research, the literature summarized the details of the domain involved and the research tools, the following part introduces some clustering algorithms in the complex network domain. This article constructs the financial network which has different significant meanings and carries on the analysis of taking the complex network technology as the foundation. By this way to find the characteristic of the unusual account intercourse network which hide under the huge normal financial network. All those works provide a kind of new research point in capital monitoring.This paper puts forward the research fruits as following:Firstly, this paper constructs a visible financial network which can describe the relationship of the account vertexes and uses a social network analysis software UCINET to find the clustering characteristic of the finance network. This may provides one kind of new point of view for the financial network monitor method.Secondly, this paper constructs a financial network which its weights mean capital transfers frequency and uses short-path algorithm to find the correlation ship of any two account vertexes, namely the most possible way of shifting capitals.Thirdly, this paper constructs a financial network which its weights mean capital transfers quantity and studies the community clustering in financial network under the theory of max flow-min cut and the theory of network community. The work can aid to find the structure of the network and proposes the certain more reliable evidence in the monitoring and controlling suspicious account vertexes.