粮食需求预测方法述评

A REVIEW ON FOOD DEMAND FORECAST METHODS

在国内外24篇有代表性的粮食需求预测文献基础上,文章重点对这些文献的预测时间、预测结果、预测方法、预测的地区、城乡和品种差异进行了比较分析。现有文献的3个主要特点是:首先,定性分析、时间序列模型和联立方程模型是最常见的三种方法,单方程模型和其他方法应用较少;第二,文献集中关注2010～2030的粮食需求预测,且在相同时间点不同文献的预测结果比较接近;第三,在预测中,少数文献关注地区差异、城乡差异或品种差异,并且同时关注3种差异的只有1篇。在对现有文献评价基础上,文章提出粮食需求预测领域应进一步深入研究的方向。

On the basis of twenty-four representative literatures about food demand forecasts, this article comparatively analyzed the forecast time, results, methods, differences among areas, urban and rural, and species. Three main features were listed as follows： first, qualitative analysis, time series models and simultaneous equations model were the most common methods, while single-equation models and other methods were seldom used; second, literatures mainly focused on the 2010 - 2030 food demand forecast and the prediction results of the same year from different paper were very similar; third, in the forecast, a small number of documents concerned regional differences, urban-rural differences or species differences, and only one concerned about the three kinds of differences together. After the evaluation on the existing literature, this article pointed out the direction of further research on the food needs of forecasting.