摘要
公路客运量预测是为了超前掌握公路客运量的发展趋势、特征和规律,有助于公路网的规划建设和管理。文章使用基于时间序列和回归分析的组合预测方法,对公路客运量进行预测。首先,分别使用单项预测方法进行客运量预测;其次,在对倒数权系数与合作对策权系数确定法进行改进的基础上,根据单项预测方法或者给予预权的组合预测方法的预测结果的平均预测误差绝对值和平均预测误差平方对预测组合的权重系数进行确定,并进行组合预测;最后,根据预测有效性判断标准,分别对比不同单预测方法之间、基于不同权系数确定方法的组合预测方法之间以及组合预测和单项预测方法之间的预测有效性。对比分析结果表明,组合预测方法预测结果的有效性较多数单项预测方法更优。
Highway passenger volume forecasting aims to know about its development trend,feature and regularity in advance and depth, and it contributes to the planning,construction and management of highway network.The paper use a combination prediction method based on time series and regression analysis to forecast the future highway passenger volume.Firstly, the paper undertakes the prediction by respectively using the single prediction method. Secondly, by analyzing the square of the average prediction error and the ABS of the average prediction error of the single prediction method and the combination prediction method,the weights of each single method of the combination are decided based on the improvement of the combination weight deciding-method with the reciprocal and the Cooperation Game,and the prediction value of the combination method can be calculated.Lastly, according to the criteria of prediction effectiveness, the paper compares respectively the effectiveness among the single method etween the combination method which adopts different combination weight deciding-method and the single method. The analysis result shows that the combination method is always a more effective forecasting method than single prediction methods.
出处
《综合运输》
2017年第2期52-61,共10页
China Transportation Review
关键词
公路客运量
组合预测
时间序列
回归分析
灰色预测
权系数
合作对策
Highway passenger volume
Combination prediction
Time Weight
Coefficient
Cooperation Game series
Regression analysis
Gray prediction
