Traffic accident characteristics analysis based on fuzzy clustering

Traffic accidents cause huge property and life losses, so it is important to study the characteristics of the traffic accidents to reduce the losses and prevent the accidents happening. Applying vectors´ inner product method of fuzzy clustering, the traffic accidents data in 2010 in China was analyzed, and the fuzzy similar matrix was established. Then the square method was applied to calculate transitive closure, and the fuzzy equivalent matrix was obtained. Based on the different values of similarity coefficient, dynamic cluster analysis was carried on. Research results indicate that comparing the traffic accident index of the five kinds of traffic modes, namely automobile, motorcycle, tractor, non-motorized vehicle, and pedestrian, traffic accident probability of automobile is the highest, traffic accident probability of tractor, non-motorized vehicle, and pedestrian is similar and in middle level, and traffic accident probability of motorcycle is the lowest.