中国棉花种植面积及皮棉产量时间序列的建模和预测研究  

Studies on Modeling and Prediction of Time Series of Cotton Area and Lint Yield in China

作　　者：王志忠[1] 王树林[1] 祁虹[1] 张谦[1] 林永增[1] 李智峰[1] 

机构地区：[1]河北省农林科学院棉花研究所,河北石家庄050051

出　　处：《河北农业科学》2011年第5期4-8,13,共6页Journal of Hebei Agricultural Sciences

基　　金：公益性行业(农业)科研专项(nyhyzx07-005);国家棉花产业技术体系项目

摘　　要：以1919～2010年中国棉花年种植面积和皮棉产量为资料,根据Box-Jenkins建模原理,采用ARIMA(p,d,q)动态模型,对中国棉花面积和产量时间序列的演变规律进行了拟合及预测研究。结果表明:所建ARIMA(2,1,2)模型的残差序列均为白噪声,AIC分别为7.878和7.426,模型结构合理,拟合效果好;2 a试报结果显示,相对精度为89.14%～99.53%,即预测准确度高。可见,运用该模型对棉花种植面积和皮棉产量的时间序列进行拟合及预测是可行的。预测2011年中国棉花种植面积及皮棉产量分别为475.89万hm2和658.58万t。该项研究为了解我国棉花生产的变化趋势及发展前景提供了1种新的研究途径。Based on the data of cotton area and the lint yield from 1919 to 2010 in China,according to Box-Jenkins theory and ARIMA（p,d,q） dynamic model,the variation law of time series of cotton area and lint yield was simulated and predicted.The results showed that the deviation of ARIMA（2,1,2） model were all the white noise,and the AIC was 7.878 and 7.426,respectively,which indicated that the model had a reasonable structure and good fittest.The relative accuracy of prediction tests for 2009 and 2010 were 89.14%-99.53%.The area and lint yield in China for 2011 were predicted as 4 758 900 hectares and 6 585 800 tons,respectively.The research offered a new way for understanding the development trend and prospect of cotton in China.

关 键 词：中国 棉花 种植面积 皮棉产量 时间序列 ARIMA模型 预测 

分 类 号：S562[农业科学—作物学]