摘要
选取城镇居民冷链食品的购买量作为冷链物流需求水平指标,尝试将基于熵权的组合预测法应用于由宁夏的冷链物流需求量的预测,以提高预测的精度。通过选用MAE、MAPE等5个预测效果的评价指标对各单一预测模型进行综合比较,利用MATLAB并结合熵权法客观地为每种预测方法分配权重。通过赋予合理权重,将各单一预测模型的预测结果加权求和得到最终的预测结果,建立了宁夏冷链物流的需求量组合预测模型,且利用2005-2014年宁夏冷链食品购买量的资料进行拟合检验,发现组合模型的精度明显提高,并预测出2017-2025年宁夏冷链物流的需求量。
In this paper, with the cold chain food purchasing value of urban residents as an index measuring the demand for cold chain logistics, we attempted to apply the entropy weighted combination forecasting process to the cold chain logistics demand of Ningxia. Through selecting five evaluation indexes of forecasting performance such as MAE and MAPE, etc., we comprehensively compared the performance of each individual forecasting model, and then using MATLAB and in connection with the entropy weight method, weighted them objectively, through which, we yielded the ultimate forecasting result and built the combination forecasting model of the cold chain logistics demand in Ningxia. Next, we had a fitness test of the model using the data of the cold chain food purchasing volume in Ningxia for the period between 2005 and 2014 to find the model was of obviously improved accuracy than its individual components and at the end, predicted the cold chain logistics demand of the area from 2017 to 2025.
出处
《物流技术》
2017年第2期117-121,共5页
Logistics Technology
关键词
熵权法
组合预测
冷链物流
宁夏
城镇居民
entropy weight method
combination forecasting
cold chain logistics
Ningxia
urban resident
