网络关键节点发现在反洗钱中的应用研究-智能与分布计算实验室

网络关键节点发现在反洗钱中的应用研究

姓名	易鑫
论文答辩日期	2007.05.31
论文提交日期	2007.06.04
论文级别	硕士
中文题名	网络关键节点发现在反洗钱中的应用研究
英文题名	An Application Research on Key Node Discovery within Network in Anti-money Laundering
导师1	李玉华
导师2
中文关键词	关键节点;重要性;反洗钱;节点?边赋权网络
英文关键词	Key node;Importance;Ati-money laundering;Node-edge weighted network
中文文摘	金融犯罪是当今业内面临的棘手问题之一，其中洗钱活动日益猖獗，严重威胁着全球经济发展和国家安全。目前的洗钱犯罪网络日益庞大、错综复杂而且形式多样，如何从洗钱犯罪网络中找出关键性节点，有效打击洗钱犯罪，已经成为反洗钱领域的一个重要课题。针对这一问题，采用基于可疑交易网络的链接发现技术，对交易网络中节点的重要性进行测度，可以有效地找出可疑交易网络中的关键节点。根据社会网络分析理论，研究节点重要性的方法主要有两类：其一是社会网络分析方法，将节点的“重要性”等价为“显著性”，并不破坏网络的整体性且通常不考虑节点的重要性；其二是节点删除的研究方法，将节点的“重要性等价为该节点被删除后对网络的破坏性”，实际上考虑的是节点删除前后图的连通状况的变化情况。从对网络破坏性的角度出发，阐述了一种无指导模式下可疑网络关键节点发现方法??基于节点删除的无权网络节点重要性测度算法。算法通过量化计算网络中每个节点删除后对网络连通状况的影响，来作为节点的重要度，从而找出网络中的关键节点。在基于节点删除的无权网络节点重要性测度算法的基础上，提出了一种节点?边赋权网络节点重要性测度算法（IMN算法）。这是一种指导模式下可疑网络关键节点发现方法，它解决了目前的研究中还未涉及的节点?边赋权网络中的节点重要性的测度问题；通过引入权值交易损失衰减系数和权值交易损失函数，使得算法有较好的扩展性和适应性。IMN算法避免了无权网络节点删除法其结果不十分可靠的缺点，在反洗钱领域知识的指导下，其结果较为准确，对有效打击洗钱活动，迅速切断洗钱网络的资金转移，具有重大的意义。根据上述方法，将IMN算法应用于可疑交易智能分析监测系统中，并通过实际数据验证了方法的有效性。
英文文摘	Financial crime is one of the thorniest issues in financial industry, because money-laundering activities are becoming increasingly rampant, it’s a serious threat to the global economic development and national security. Currently money-laundering network is increasingly large and complex and diverse, how to identify the critical nodes of money-laundering networks, effectively combating money-laundering crime, anti-money laundering has become a major topic. In response to this problem, through link discovery technology based on suspicious transactions network, suspicious transactions network nodes importance measure to be effective in identifying suspicious transactions of key network nodes. Social network analysis theory, the method that researching importance of the nodes there are two main categories: First is the social network analysis method that the node will "importance" equivalent to "significant" will not undermine the integrity of the network and usually do not consider the importance of the node sets; The second is to delete nodes research methods, nodes will "importance" equivalent to "destruction after the deletion of nodes on the network", actually consider the plan before and after the deletion of nodes connected to the changes in conditions. Based on the network destructive perspective, presents a suspicious key network nodes link discovery without guidance: Importance Measurement for the Node within Non-weighted Network Algorithm base on node removed method. Quantification of the calculation the impact of network connectivity status after a network node removed, as an important node, in order to identify the key network nodes. Based on Network Node Important Measurement Algorithm Base on Node Removed Method, presents Importance Measurement for the Node within Node-edge Weighted Network Algorithm (IMN Algorithm). This is suspicious network key nodes link discovery with guidance, It solves the current research has not yet come to the node-edge weighted network nodes importance of the measure; By introducing weights transaction losses attenuation coefficient and the weights transactions loss function, Algorithm makes a better scalability and adaptability. IMN Algorithm avoids that the results of node within non-weighted network removed is not very reliable, it’s more accurate under the guidance of the field knowledge. Under this approach, we applied IMN Algorithm to analysis of suspicious transactions intelligent monitoring system, and validated the effectiveness of the method through the actual data.