犯罪信息网络分析与建模

基于2012年美国数学建模C题的数据,利用83人共15个种类的600条信息进行了犯罪信息网络分析,建立了概率网络模型和最短路径模型,设计了相应的求解算法,对所有人的可疑度进行评价和排序,找出通信网络中的可疑嫌犯,并对两个模型进行了对比。然后基于中心性理论建立了识别嫌犯领导人的模型,得到犯罪集团中最可能的领导人。最后讨论了文本分析、语义网络分析方法在犯罪信息网络分析中的应用,并对模型在其他领域推广应用的可行性进行了探讨。

This paper is about network analysis of criminal message based on the data of 2012 ICM. Using 400 messages of 15 themes among 83 people, we establish two models and corresponding algorithms to calculate the likelihood of one’s being conspirator (probability of a node) respectively based on probability network analysis and graph theory, then sort them to give a list of probable conspirators and make a comparison of two models. We recognize the possible leaders of conspirators by applying centrality theory. At last, we discuss the application of text analysis and semantic network analysis in criminal information network analysis, and application in other disciplines of our models.