论文部分内容阅读
为提高交通事故预测精度,基于熵值法构建UGM(1,1)-Holt组合预测模型,将滑动转移概率思想引入马尔科夫链模型,实现组合预测优化。应用该模型拟合分析2003—2011年湖北省交通事故死亡人数的历史数据,并以2012—2014年数据验证其有效性。通过实例对比UGM(1,1)模型、Holt指数平滑模型、组合预测模型和组合预测优化模型的预测精度。结果表明:相比前3种模型,提出的组合预测优化方法拟合值平均相对误差(MRE)为0.45%,3年预测值MRE为1.25%,能有效获取单一模型优势,预测精度更高。
In order to improve the accuracy of traffic accident prediction, UGM (1,1) -Holt combination forecasting model is constructed based on entropy method, and the idea of sliding transition probability is introduced into Markov chain model to realize the combined forecasting optimization. The model was used to fit and analyze the historical data of traffic fatalities in Hubei province from 2003 to 2011. The data of 2012-2014 were used to verify the validity of the model. The precision of prediction of UGM (1,1) model, Holt index smoothing model, combined forecasting model and combined forecasting optimization model are compared by examples. The results show that compared with the former three models, the combined forecasting optimization method proposed has an average relative error (MRE) of 0.45% and a 3-year forecast of MRE of 1.25%, which can effectively capture the advantages of a single model and achieve higher prediction accuracy.