摘要
为探究我国粮食价格波动规律及预测其值,选取2005-2016年中国各季度粮食生产价格数据,首先对其波动性进行描述分析;其次建立一种基于IGOWLA算子的组合预测模型:选择ARMA模型、Holt-Winters乘法模型、残差自回归模型这三种单项预测模型,并通过5种误差评价指标来判断预测模型的效果,结果表明组合预测模型的预测效果较好、准确性较高;接着利用所建立的组合预测模型对2017年各季度粮价进行外推预测,结果表明2017年粮食价格有所上升但相对波动较稳;最后提出政策建议。
In order to explore the law of Chinese grain price fluctuation and its influencing factors, first we choose quarterly grain price data from 2005 to 2016 and analysis it. Then, a combined prediction method based on the induced generalized ordered weighted logarithmic averaging operator is proposed. We use the ARMA model, the Holt-Winters multiplication model and the residual autoregressive model respectively to make an individual prediction of Chinese grain production price. Then, we establish the combination forecasting model based on IGOWLA operator. By comparison, it is found that the latter has higher accuracy and stronger predictive ability than the former.Next, we obtain the weights under the basement of the IGOWLA operator and make predictions and analysis to the grain production price from the first quarter of 2017 to the fourth quarter of 2017.Last, we make some suggestions.
出处
《怀化学院学报》
2017年第4期31-36,共6页
Journal of Huaihua University
基金
国家社科基金项目"组合预测模型与方法创新及其优化理论研究"(12BTJ008)
安徽财经大学研究生科研创新基金项目(ACYC2016115)
