2012年美国建模比赛提交作品.pdf

Team #13553 Page 1 of 15 Model for crime busting Summary Using the principal component analysis and factor analysis approach, a simple numerical model is raised to solve a criminal case. We analyze the conversations between people, and use the software SPSS (Statistical Product and Service Solutions) to count each person’s score which represents the likelihood of one to be a conspirator. Then we rank those scores so that we can prioritize people by the likelihood of being part of the conspiracy. Under our model and algorithms, we can well judge whether one is likely to be involved in the conspiracy. Besides, our model can be applied to analyzing other cases such as analyzing a biological network‐like database. Key words: principle component analysis, factor analysis, text analysis, likelihood of crime. Team #13553 Page 2 of 15 1. Introduction 3 1.1 Problem restatement 3 1.2 Problem analysis 3 1.3 Assumptions 4 2. Model for the crime busting 5 2.1 Requirement 1 5 2.1.1 Categories classification 5 2.1.2 Count the frequency 5 2.1.3 Principal component analysis and factor analysis 6 2.2 Requirement 2 12 2.3 Requirement 3 13 2.4 Requirement 4 14 3. References 15 Team #13553 Page 3 of 15 1. Introduction 1.1 Problem restatement The Intergalactic Crime Modelers (ICM), by using a network‐like database, is investigating a conspiracy. There are 83 nodes in the network‐like database which represent 83 people in the company, 400 links which represent the conversations between every two people, and totally 15 topics (the topic7 ,11 ,13 have been deemed to be suspicious). At the same time, it is known that Jean, Alex, Elsie, Paul, Ulf, Yao, and Harvey are conspirators. Also, it is known that Darlene, Tran, Jia, Ellin, Gard, Chris, Paige, and Este are not conspirators. A prio