The evaluation method of road traffic safety based on the combined Principal Component Cluster analysis

摘要

On purpose of scientifically and reasonably evaluating the present state of traffic safety, the improved Principal Component analysis and Cluster analysis are combined and the deficiency of existence of negative variation coefficient in Principal Component analysis could be made up by value assignment analysis. First of all, the synthesis evaluation indicators system is constructed based on nine relative indicators such as mortality rate per ten thousand vehicles, injury rate per ten thousand vehicles, mortality rate per million people, injury rate per million people, crash frequency per mile, extra serious accident frequency per mile and so on, which could comprehensively reflect the safety of road traffic. Then, the data from the eastern provinces of China in 2009 are synthetically evaluated step by step as an example by applying the traditional and improved Principal Component analysis methods. Lastly, the evaluation results are compared and analyzed according to the differences between these two methods by the Cluster analysis method.