高速公路交通量预测的GM(1,1)残差改进模型

针对传统GM(1,1)模型在预测高速公路交通量中存在误差过大、计算复杂的问题,通过定义残差序列,对预测序列与残差序列进行累加再处理,构造新的序列数据;并且对新序列数据构造GM(1,1)残差改进模型,以进行预测。模型应用于某高速公路某收费站,对9期序列数据进行了模拟预测。结果表明,GM(1,1)残差改进模型的平均预测误差为7.25%,优于传统GM(1,1)模型预测的平均相对误差12.7%。

Residuals improved GM(1,1) model of expressway traffic volume prediction

Expressway traffic has typical characteristics of gray system.In order to minis the excessive error and computational complexity from traditional GM(1,1) model in expressway traffic volume prediction,this paper defines the residual series,formulates a new sequence by using the sum of predicting sequence and the absolute value of residual sequence so as to construct a residuals improved GM(1,1) model to predict expressway traffic volume.The model is applied to an expressway for predicting 9 series data.The results shows that the average error of residuals improved GM(1,1) model is 7.25%,which is significantly better than that 12.7% of traditional GM(1,1) model.3 tabs,9 refs.