2011国际赛lzx.docVIP

For office use only T1 ________________ T2 ________________ T3 ________________ T4 ________________ Team Control Number16130 ? Problem ChosenC For office use only F1 ________________ F2 ________________ F3 ________________ F4 ________________ A Report on 2012 ICM Problem Modeling for Crime Busting Abstract Crime Network Analysis(CAN)，a form of Social Network Analysis(SNA), is specialized analysis of social nodes which take the forms of members of a criminal conspiracy and other people related to them. We offer two models for the current case: a Semantic Network Analysis Model and a Criminal Network Model. The former is built to calculate the relevance of each topic to the conspiracy, and the later works on the relevance of each person to the conspiracy. Firstly, in the Semantic Network Analysis Model, we use Text Analysis to single out the key words in the known suspicious topics, and then classify them and define their weights（0~1）. We match the key words in the other uncertain topics and calculate their weights. In particular, we use arc-tangent function repeatedly to simulate the pattern in which the value of weights increase, and work out the weights(similarity) between network nodes. With a reference to IDW, we calculate an adjacency matrix of equivalent distance between nodes in the criminal networks. Secondly, in the Criminal Network Model, we use Network Centrality, which includes Degree, Closeness, and Betweenness, to measure and analyze the nodes in the network. By calculating Closeness we get a priority list of each person’s relevance to the conspiracy. By calculating Betweenness we identify Alex as the leader of the conspirators. Taking into account the conditions in Requirement 2, we notice that the priority list remains much the same. Besides, we adopt RNN-CLINK Cluster Analysis to further analyze the structure in the gang, and identify each of its sub-gang. At the same time, reliability of the results of calculation is testified. Finally, we