摘要
针对单一预测方法存在的不足,采用一种联合时间序列法和因果分析法的混合算法预测港口吞吐量.并以宿迁港吞吐量预测为例,对这种方法进行了详细说明.预测过程中借助主成分分析法提取出影响港口吞吐量的关键因子,利用改进的灰色模型理论对关键因子中远期未来值做出预估,最后通过吞吐量与关键因子的线性回归得出宿迁港在预测年份的吞吐量值.通过对混合算法预测模型误差的检验发现,此预测方法的模型精度较高,能够满足港口吞吐量预测的要求,可以推广到类似港口吞吐量的预测中.
In view of disadvantages of single prediction method, a hybrid algorithm which combines Time Series Method and Causality Analysis is used to predict the port through- put, and the Suqian port is taken as an example for detailed use of the method. In the process of prediction, the principal component analysis is used to extract critical impact fac- tors, and improved grey model is used to predict medium and long term value of the critical impact factors in the future. By the analysis of linear regression between throughput and critical factors to get the throughput prediction of Suqian port in forecast years. By the test of errors in this prediction model, the paper indicates the prediction accuracy of the model is quite high, which can meet the demand of port throughput prediction, and can be general- ized to other ports＇ throughput prediction.
出处
《长沙理工大学学报（自然科学版）》
CAS
2012年第4期47-52,共6页
Journal of Changsha University of Science and Technology:Natural Science
关键词
港口吞吐量
混合算法
主成分分析法
改进的灰色模型
port throughput
hybrid algorithm
principal component analysis
improved grey model
