Research on suspicious transaction recognition based on heuristic listing of directed primary circuit
Xu Taihua1*，Zhang Qinghua2
Author information+
{{custom_zuoZheDiZhi}}
About authors:
1.School of Information Science and Technology，Southwest Jiaotong University，Chengdu，610031，China；2．Chongqing Key Laboratory of Computational Intelligence，Chongqing University of Posts and Telecommunications，Chongqing，400065，China
An important topic of antimoney laundering is recognition of suspicious financial transaction.Directed primary circuit(DPC)is a very important structure in graph.Financial transaction data can be expressed by directed graphs，called financial transaction graphs，and it is obviously that DPC is suspicious if it exists in financial transaction graphs，because it means that money may return to starter.In this paper，we design a heuristic listing of DPC algorithm to help us to recognize these suspicious structures.The basic idea of the algorithm is to calculate strongly connected components(SCC)by using Tarjan’s strongly connected components algorithm at first，and then to conduct heuristic depthfirst search(DFS)on every SCC which has been found.Differ to normal DFS，in the process of heuristic DFS，two heuristic information are utilized，one is parameter of direction to control the direction of DFS，and the other is indegree and outdegree in SCC that are newly defined to choose visiting vertex by descending indegree or outdegree.Also，we prove that there must be a DPC in a SCC which has 3 vertexes at least，and this proof provides theory basis to proposed algorithm.Certainly the time complexity of proposed algorithm is analyzed.Proposed algorithm reduces the size of query space，which leads to enhancement of algorithm performance.In order to test proposed algorithm，we choose several datasets of University of Florida Matrix Collection，and collect several practical datasets.All of datasets can be transformed to directed graphs.We design two experiments，one is for testing effectiveness of proposed algorithm，and the other is for validating necessity of two heuristic information.Experimental results show that our algorithm is effective，and utilization of two heuristic information is necessary.Thus，suspicious financial transactions，DPC，can be listed in financial transaction graphs，and technical assistance can be provided for antimoney laundering.
Xu Taihua1*，Zhang Qinghua2.
Research on suspicious transaction recognition based on heuristic listing of directed primary circuit[J]. Journal of Nanjing University(Natural Sciences), 2016, 52(5): 879
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] Tang J，Yin J．Developing an intelligent data discriminating system of antimoney laundering based on SVM.In：Proceedings of 4th International Conference on Machine Learning and Cybernetics.Guangzhou：IEEE Press，2005：3453－3457. [2] Zhang Z，Salerno J J，Yu P S．Applying data mining in investigating money laundering crimes.In：Proceedings of 9th International Conference on Knowledge Discovery and Data Mining.Washington DC：ACM Press，2003：747－752. [3] Brabazon A，O’Neill M．Biologically Inspired Algorithms for Financial Modelling.Genetic Programming & Evolvable Machines，2006，7(3)：285－286. [4] Yang S，Wei L．Detecting money laundering using filtering techniques：A multiplecriteria index.Journal of Economic Policy Reform，2010，13(2)：159－178. [5] 杨冬梅，吴冲锋，冯 芸．金融网络中洗钱资金异常转移路径的经济成本模型．系统工程理论与实践，2006，26(5)：23－29.(Yang D M，Wu C F，Feng Y．The economiccost model about laundering money abnormal transfer paths in the financial network.System EngineeringTheory & Practice，2006，26(5)：23－29.) [6] 薛耀文，王雪娟，张鹏翥．洗钱交易模式及其防范分析．系统管理学报，2008，17(4)：418－422.(Xue Y W，Wang X J，Zhang P Z．Analysis on the model of money laundering and prevention.Journal of System & Management，2008，17(4)：418－422.) [7] 张成虎，高 薇．基于朴素贝叶斯分类的可疑金融交易识别研究．情报杂志，2006，25(11)：46－47.(Zhang C H，Gao W．Research on suspicious financial transaction recognition based on naive bayesian classification.Journal of Information，2006，25(11)：46－47.) [8] 张成虎，赵小虎．基于链接分析的洗钱交易识别研究．上海金融，2009(8)：78－83.(Zhang C H，Zhao X H．Research on identify money laundering transaction based on link mining.Shanghai Finance，2009(8)：78－83.) [9] 张成虎，尹 为，杨 彬．基于数据流频繁子图挖掘的可疑金融交易动态识别．系统工程，2013(7)：1－7.(Zhang C H，Yin W，Yang B．Research on suspicious financial transactions recognition based on frequent item digging in stream data.Journal of Xi’an Jiaotong University(Social Sciences)，2013(7)：1－7.) [10] 尹 为，张成虎，甘 凯．基于数据流多维分析的可疑金融交易动态识别．北京理工大学学报：社会科学版，2013，15(5)：52－59.(Yin W，Zhang C H，Gan K．Dynamic suspicious financial transactions recognition based on multidimension analysis of data stream.Journal of Beijing Institute of Technology(Social Sciences)，2013，15(5)：52－59.) [11] 喻 炜，王建东．基于交易网络特征向量中心度量的可疑洗钱识别系统．计算机应用，2009，29(9)：2581－2585.(Yu W，Wang J D．Suspicious money laundering detection system based on eigenvector centrality measure of transaction network.Journal of Computer Applications，2009，29(9)：2581－2585.) [12] Tarjan R．Enumeration of the elementary circuits of a directed graph.SIAM Journal on Computing，1973，2(3)：211－216. [13] Johnson D B．Finding all the elementary circuits of a directed graph.SIAM Journal on Computing，1975，4(1)：77－84. [14] Raina R R，Katare R K，Mughal S A，et al.An optimization algorithm for finding graph circuits and shortest cyclic paths in directed/undirected graphs.International Journal of Scientific and Engineering Research，2014，5(1)：472－476. [15] Liu H，Wang J．A new way to enumerate cycles in graph.In：Proceedings of Advanced International Conference on Telecommunications and International Conference on Internet and Web Applications and Services(AICT/ICIW 2006)．Guadeloupe，French Caribbean：IEEE Computer Society Press，2006：57－57. [16] Tarjan R．Depthfirst search and linear graph algorithms.SIAM Journal on Computing，1972，1(2)：146－160. [17] Ooof L．Gephi chinese tutorial.https：//www.udemy.com/gephi/.2012. [18] Batagelj V，Mrvar A．Pajek datasets.http：//vlado.fmf.uni－lj.si/pub/networks/data/.2006.
PDF(3551938 KB)
