基于向前选择变量法的我国粮食总产量多元线性回归预测模型

Multiple linear regression forecasting model of total food yield in China based on forward selection variables method

结合1983～2006年我国粮食总产量及与之密切相关的粮食单产、粮食播种面积、化肥施用量、有效灌溉面积、农机总动力、农田成灾面积等6个影响因子序列资料,采用向前选择变量法,构建了我国中长期粮食总产量多元线性回归预测模型。结果表明,粮食单产、粮食播种面积、化肥施用量及农田成灾面积是我国粮食总产量的关键制约因子,措施得当,未来我国完全可以在耕地资源不可逆转减少的前提下实现粮食总产量的持续增长。研究成果可为我国粮食安全保障体系的构建提供决策依据。

The mid and long term total food yield multiple linear regression forecasting model of China was set up through using the forward selection variables method and combing with the series data of Chinese total food yield and it＇s closely related six influencing factors from 1983 to 2006,which were food yield per unit area,grain planting area,fertilizer application rate,effective irrigation area,total power of agricultural machinery and destroyed farmland area.The results showed that food yield per unit area,grain planting area,fertilizer application rate and destroyed farmland area are the key limiting factors of Chinese total food yield,and China could fully realize the continuous increase of total food yield under the premise of plantation resource irreversible decrease in the future,as long as the measures were suitable.The research achievements can provide decision basis for the establishment of food safety guarantee system in China.