Research on Grey-Markov prediction model and self-adaptive improvement for traffic accidents due to information deficiency

As a joint-effect result of driver, vehicle, road and environment in the complex traffic system, the regularity and transfer direction of traffic accidents displays the representative character of grey and information deficiency. On the basis of analyzing the grey character of traffic system, a data-disposal tool, grey model, constructed by least-square theory, is adopted to prejudge the trend of development and future behavior of traffic accidents. A as a amendment factor of boundary value, Markov-chain and self-adaptive manipulation are introduced to improve the prediction precision and lessen the effect of data fluctuation responding to its deficiency of extrapolation ability. An evaluation example proves SAGMM prediction reveals a high-accuracy performance, overmatches single grey model, practises programmed disposal with mass data and possesses the value of practicality in traffic administration and policy legislation in response to data series of information uncertainty and fluctuation.