摘要
交通量具有高度复杂的非线性特征,采用单一预测模型往往难以达到理想的预测效果。为准确预测,提出一种最优线性组合预测模型并给出了以预测误差平方和最小为目标函数的权系数最优解计算方法,在采用ARIMA模型、BP神经网络和支持向量回归机的基础上,利用组合预测模型实现了高速公路月度交通量的预测。实验结果表明:与季节差分自回归滑动平均模型、BP神经网络和支持向量回归机等预测模型相比,组合预测模型各项评价指标均优于前三者,为实现交通量准确预测提供了更为科学的依据。
ABSTRACT： As having high complex nonlinear characteristic, the prediction effect of traffic volume is usually unsat- isfactory. On the basis of seasonal ARIMA model, BP neural network and support vector regression, this paper pro- posed an optimal linear combination prediction model and presented a method for calculating optimal solution of weight coefficients, taking the minimized sum of squared errors as objective function. The combinational model real- ized the prediction of expressway monthly traffic volume. The experimental results show that evaluation indices of the combination prediction model are better than seasonal ARIMA model, BP neural network and support vector regres- sion. It provides a more scientific method for realizing accurate prediction of traffic volume.
出处
《计算机仿真》
CSCD
北大核心
2013年第4期178-182,共5页
Computer Simulation
基金
国家自然科学基金项目(60804049)
教育部创新团队发展计划资助项目(IRT1050)
