Analysis of Traffic Accident Based on Fractal Theory
Peng Chen, Xuhong Li, Sun Hua-Can · Journal of Highway and Transportation Research and Development · 2008
Taking statistic data of traffic accidents of a using fractal theory was demonstrated firstly by power city as an example, the feasibility to analyze traffic accident spectrum analysis. The calculation result shows that in the sense of statistics, time series of annual distribution of traffic accidents has certain self-affinity. Then the fractal property of inner interval was extended outwards by self-similarity and scale invarianee of fraetal, and on this basis a fraetal extrapolation algorithm was proposed to realize the prediction of traffic accidents. By this algorithm the search was performed from a certain initial point by use of iterated function system and attractor of inner interval. When the mean square deviation between point set obtained by iteration and attractor become the minimum, the prediction value can be used as the function value of the point to be extrapolated. Finally, the R/S fractal analysis of time series of traffic accidents further shows that the forecasting value is reasonable. Case study results show that there will still be growth trend in the next few years and traffic management should be continued to strengthen.