| With the rapid growth in the number of China’s motor vehicles and the extension of road mileage, traffic pressure has greatly increased and the traffic problems have become increasingly serious, making the research in traffic problems vitally significant. Although the research in traffic problems can be conducted from different angles, this paper tries to deal with them from the aspect of avoiding traffic violations which are the main cause of traffic accidents. Due to the great value of traffic violation data, it is necessary to conduct comprehensive data analysis and data mining. Spatial co-location pattern mining is applied to study traffic violations in this paper in order to provide scientific data to support government decisions in making traffic laws, and to provide good consultancy for the prevention of traffic accidents and the guarantee of clear roads.Based on the spatial co-location pattern mining method, we analyzed the spatial relationships between the characteristics of traffic violations and road design features, gas stations, hospitals, schools, and bus stations. Considering the large data of traffic violations, we adopted a simulative data on illegal traffic behaviors in March 2013 in a city, and intercepted part of the accidents in a certain range; then, various types of traffic violations and road designs were sorted out by hand; The coordinates of extracted accidents were preprocessed, with the spots being presented as latitudes and longitudes. As a certain latitude and longitude was seen as the original coordinate, latitude and longitude values could be converted into the length values for easier computation of adjacent spatial relations. By using traditional join-based algorithm for mining spatial co-location patterns in.NET environment, the traffic violation data mining was carried out. Finally, based on the co-location rule mining, we analyzed the practical significance and value of the mined results, in order to provide relevant advice for preventing traffic violations and accidents. |
