摘要
在分析现有货运量预测方法所存在问题的基础上,建立了货运量预测的支持向量机模型,并以我国1981～2001年的货运量和相关经济指标的历史统计数据作为学习样本,分别选用3种不同的核函数,通过拟合训练和外推预测分析,验证了支持向量机用于货运量预测的有效性,并对模型中的有关参数进行了探讨分析。
The paper analyzes some existing proldems of the present freight volume forecast methods. Then the paper establishes the support vector machine model for freight volume forecasting. Using the historical statistical data of freight volumes and related economic indices from 1981 to 2001 as and the learning specimens and choosing three different kinds of kernel functions to do fit training and extrapolation forecast analysis, the paper demonstratesd the effectiveness of using the support vector machine in freight volume forecast. At last, the paper analyzesd some parameters of SVM.
出处
《铁道学报》
EI
CAS
CSCD
北大核心
2004年第4期10-14,共5页
Journal of the China Railway Society
