MCM ICM优秀论文剖析.ppt

In addition, we make the assumption that sending and receiving messages have same effect when we evaluate the likelihood of one’s being conspirator; 另外我们在确定一个人是同谋者的可能性的大小时，假定发送和接受信息是等效的。 Based on psychology analysis above, we can state that all conspirators have at least one message communication with other conspirators, whether suspicious or unsuspicious message. 基于上面的分析，我们认为所有的同谋者至少有一次与其他同谋者的信息交流，不论是否是可疑信息的交流。 Model 1 Establishment of model （简单统计建模） To evaluate the likelihood of one’s being conspirators, we use the following equation which combines two quantity indexes:某人是同案犯的可能性度量 用下面的公式计算： Where is the suspicious message number that office worker i sent or received and is message number that office worker i sent to or received by known conspirators. 这里 是办公室人员i 发送或接受可疑信息的数目。 是办公室人员 i 与已知的同谋者之间的发送或接受的信息数目。 Figure. 1 Result and analysis Figure 1 shows all the values of and . Using equation (1) we have put forward, we can easily calculate all the values of and make a priority list as Table 1 (note that is not a probability but a metric to evaluate the likelihood, though it value is between 0 and 1) 图1 给出了所有的 和 ，用方程（1）可以计算出 . 不是概率，是一种度量） As shown in Table 1, all the known conspirators (heavy tape and red mark) are ranked in the very front of the list, which indicates the model is effective to some extent that it can recognize some workers who is most likely to be conspirators. However, some non-conspirators (green mark and Italic type) are also up at the front, like node 48 and node 2, which shows that the model has a certain limitation and some wrong recognition. 表1中所有已知的同谋者（红色粗体）都在表的最前面，说明了模型的有效性。它能识别最可疑的同谋者。但是一些确定的非同谋者也在表的前列，例如48和2，这说明这个模型的局限性和存在错误的识别。 Model 2 网络—概率模型 more assumptions： Except for the known conspirators and non-conspirators, one’s probability of being conspirator is relate to those who have direct message contact with him/her. And the probability is both affected by the probability o