【作者】 李亮；

【导师】 曹健； 张虹；

【作者基本信息】 上海交通大学 ， 软件工程， 2008， 硕士

【摘要】 随着社会经济的快速发展,团伙犯罪的发案率不断升高,相对于一般刑事案件,团伙犯罪的社会危害性更大。如何遏制和打击团伙犯罪已经成为各地公安机关当前工作的重点任务。本论文从公安实际工作的现状和需要出发,将社会网络分析的相关理论与方法引入到公安部门对犯罪团伙的识别工作,建立了一个符合公安工作实际需求,基于已掌握犯罪团伙某一个或多个嫌疑人资料的情况下,缩小团伙中其余人员排查范围的辅助系统,使之成为又一个打击犯罪、维护社会治安、保护人民生命财产安全的有力武器。本文将工作重点放在以下几方面:1.根据公安各部门现有业务数据库的特点,对已经掌握的嫌疑人各社会关系的相关数据进行清洗、转换,并建立相应的案件数据仓库,建立该嫌疑人的社会网络;2.阐述了多种社会网络中社团的分解算法,利用Radicchi等人的快速分裂算法对已建立的嫌疑人社会网络进行分解,缩小可能的犯罪团伙排查范围;3.根据复杂网络中的聚类算法,利用基于SPLINE(Shortest Pathalgorithm Based on Link Weight)的犯罪集团网络核心挖掘算法KMM(Key-Member Mining)对各子网进行社会网络分析,突出可能的犯罪团伙中的核心人员。4.在上述具体技术基础上,我们进行了犯罪团伙识别的设计和实现。其中主要包括了系统整体框架设计、主要功能模块的设计。最终实现并构建了整个系统。最后,我们根据已侦破的一起团伙犯罪案件的相关情况,进行了基于社会网络的犯罪团伙识别模拟测试。测试结果表明,该系统对犯罪团伙的识别筛选出的参考结果,相对于以往实际办案中进行排查的范围来说,已经有了明显的缩小,对这个筛选结果再进行很小工作量的人工排查就可以得到满意的结果。从而提高工作效率,降低工作强度,达到辅助决策的目的。本文做了以下创新工作:1.实现了犯罪团伙自动识别技术在苏州市公安局的应用。犯罪团伙的自动识别在我国公安机关的使用尚属起步阶段,在苏州市公安局范围内的应用更属首次;2.将公安机关各部门现有的业务数据库进行了集中整合,建立了面向案件的数据仓库,极大地提高了原业务数据的利用效率。更多还原

【Abstract】 With the rapid development of economy, the criminal rate about criminal group is increasing. The criminal committed by the criminal group is more dangerous than the common criminal. It is an important work to public security that how to fight against the criminal group.In this paper, we analyze the work of current situation and the requirement of the public security. We use the social network analysis to help the public security recognize the criminal group and form an auxiliary system for the public security based on some of the members of the criminal group. The system can help the public security find the unknown members of the criminal group. It can be a powerful weapon to fight against the criminal, protect people’s life and property and maintain law and order.The main works are described as follows:According to the databases of the public security, clean and transform the data about the criminal’s social relations. Form the data warehouse of the cases and social network of the criminals.Talk about several kinds of association decomposition algorithm for social network. Use Radicchi’s quick split algorithm to decompose the social network for reducing the scope of investigation.According to the cluster algorithm for complex network, use the criminal group network’s KMM algorithm based on SPLINE(Shortest Pathalgorithm Based on Link Weight)to analyze all sub-networks and find the core staffs of the group.Based on the technology mentioned above, we design and implement the Recognition System of Criminal Group Based on Social Network Analysis. This includes system structure design, functional module division and database design. Then we implement them and build the system.At the end of this paper, we test the system with an uncovered criminal group case. The test result shows that the system can reduce the investigation scope greatly. After screening the result of the system manually, we can get the satisfactory result. So the system did improve efficiency, reduce the intensity of the work and assist to make a decision. The innovation works are described as follows:(1) We fulfill criminal group recognition in Suzhou Public Security Bureau. In our country the use of criminal group recognition in Public Security is still in the initial stage. And it is the first use in Suzhou Public Security Bureau.(2)Integrate all departments’databases of public security. Form the case-oriented data warehouse and improve the efficiency of data use.更多还原