摘要
短期货运量的预测对于交通运输系统的运作与管理具有重要意义。本文将回归分析与时间序列分析相结合,提出一个带有回归项和时间序列误差项的回归-时序混合模型,用以进行短期货运量的预测。在对模型进行识别、初估计的基础上,采用极大似然方法进行参数估计,经反复拟合,并对模型进行相应检验,最后得到符合要求的拟合模型。应用此回归-时序混合模型进行月度货运量的拟合预测,并与多元线性回归模型和季节ARIMA模型的拟合预测结果相比较,表明回归-时序混合模型可以提高短期货运量的预测精度。
Short-term freight prediction is important for us to manage the transportation system. Integrating regression analysis with time series analysis, aregression model with seasonal ARIMA errors -- Regression-Time Series Analysis model--was presented to forecast the short-term freight. After a series of processes of identifying, estimating with ML, fitting and testing. This paper applied the regression-time series analysis model to predict monthly freight, and compared the result with those from the multivariate linear regression model and seasonal ARIMA model. Experimental results show that the regression-time series analysis model can improve the precision in short-term freight prediction.
出处
《交通运输工程与信息学报》
2008年第1期65-68,85,共5页
Journal of Transportation Engineering and Information
关键词
货运量
预测
回归分析
时间序列分析
季节ARIMA
极大似然
Freight, prediction, regressionanalysis, time series analysis, seasonal ARIMA,maximum likelihood
