Application Of Association Analysis In Mining Of Escort Vehicle Involved In Crime Case

With motor vehicles keep growing in urban, more and more corresponding problems come along. In which, Vehicle theft and criminal activities by using of vehicles is deserve to be noticed especially. Considering Vehicle involved in Crime Case (VCC), because of motor vehicles with high mobility, concealment and other characteristics, it have characteristics of easy to escape the scene quickly and insusceptible to the investigation. It becomes an important factor that is serious threat to social security. Therefore, effective monitoring and management of motor vehicles become a new interesting and concerned problem. Through a large number of researches and analysis of crime cases involved motor vehicle in past, the involved motor vehicles are often found frequently in the vicinity of the crime scene and often with some escort vehicles.Based on the project, Shenzhen Grid-enabled Vehicle Identification Comprehensive Application System, with the vehicles licensing data identified by the initial operation of the system, escort vehicles, historical route and candidates of VCC are analyzed within the monitoring region. The identification data is huge, which reach by a few million everyday only in 80 monitoring points of the first phrase projects. Thus, how gain escort vehicles, historical routes and candidates of VCC fast and efficiently become problems that are worthy mostly of study.In recent years, various methods of data mining are got matured and applied widely in various fields. It provides some good methods and ideas to solve above problems. Based on traditional Boolean association analysis, escort vehicles, historical routes of VCC are studied. And for the problem of picking out escort vehicles, a new Extended Boolean Association Analysis (EBAA) based on the similarity of the association, which is an improvement of traditional Boolean Association Analysis (BAA), is put forward and presented. The results show that EBAA is more effective than tradition BAA by choosing the appropriate similarity and the threshold of rule-extraction.The paper mainly includes the following five points:Analyze the background of the task, determine research objectives and abstract issues and gained its mathematical description.On the basis of review of data mining methods, focus on the research progress of association rule mining method and its applications in the field of transportation. Aiming at above problems, extract, convert and clean the vehicle license data to be available to the Boolean association analysis.Analyze historical routes, escort vehicles, potential candidates of VCC by using Boolean association analysis method.Put forward the EBAA method, and use it to analyze escort vehicles, potential candidates of VCC. Through making comparison between results of BAA and EBAA, it shows that EBAA is more effective than traditional BAA.