摘要
针对影响农产品物流需求的各种特定因素,提出基于偏最小二乘回归(PLS)、BP神经网络和ARIMA模型的我国农产品物流需求量最优组合预测分析方法体系,并基于1995-2015年中国农产品物流需求量相关数据对我国农产品物流需求进行预测分析。实证分析结果表明,预测方法不但能够有效预测农产品物流需求量的整体变化趋势和细节波动,具有更高的预测精度,而且能够反映粮食产量、粮食价格等影响因素对物流需求量的作用机制。
The effective forecast of agricultural product logistics demand has importantly theoretical and practical significance to the rational planning of agricultural product logistics and improving the efficiency of agricultural product logistics to meet the needs of economic development. In this paper, an optimal combination forecasting framework of agricultural product logistics demand in China based on partial least squares regression （PLS） , BP neural network and ARIMA model is proposed. Based on the relevant data of agricultural products from 1995 to 2015, China＇s agricuhural product logistics demand using the proposed forecasting framework is forecasted and analyzed. The results of the empirical analysis show that the forecasting method can not only forecast the overall trend and detail fluctuation of the logistics demand of agricultural products, but also forecast the effect of grain yield and grain price on the agricultural product logistics demand.
出处
《佳木斯大学学报（自然科学版）》
CAS
2017年第5期852-856,共5页
Journal of Jiamusi University：Natural Science Edition
基金
国家自然科学基金(71501002,71371011,71301001)
教育部人文社会科学研究青年基金(13YJC630092)
安徽省自然科学基金(1508085QG149,1608085QF133)
安徽省社科规划项目(AHSKQ2014D13,AHSKQ2016D13)
安徽大学国家级大学生创新训练计划项目(201610357093)
