作　　者： (王星）; (刘小勇）;

机构地区： 新疆农业大学机械交通学院,新疆乌鲁木齐830000

出　　处： 《交通科技与经济》 2017年第4期9-13,共5页

摘　　要： 以全国交通事故数为研究对象,旨在实现对道路交通事故的有效预测,通过采用灰色模型和马尔科夫模型相结合的方式,建立灰色马尔科夫模型。依据2005~2014年的全国交通事故起数对模型进行精度验证,并对未来3年的交通事故数进行实例预测。结果表明:灰色马尔科夫模型的残差均值相比单一的灰色模型降低了0.020 4,平均相对误差降低了4.45%,预测精度有明显提高,预测结果具有动态波动性,更符合实际需求,从而为道路安全管理提供决策支持。 In order to realize the effective prediction of road traffic accidents, the Gray model and Markov model are combined to a Gray Markov model. The accuracy of this model is verified by the national traffic accidents from 2005 to 2014, and the number of traffic accidents is forecasted for the next three years. The result shows that the residual mean of the Gray Markov model is reduced by 0. 020 4 and the relative average error is reduced by 4.45% compared with the single gray model. In addition, the prediction accuracy is obviously improved and the forecasting result has dynamic volatility, which accords with the actual demand, thus providing a decision support for the road safety management.

关 键 词： 道路交通事故 灰色模型 灰色马尔科夫模型 动态波动性 决策支持