1、并且其白噪声检验结果如下,证明信息已经被完全提取。最后得出结果模型为则去掉差分后的ARIMA(2,2,0)模型为:利用该模型预测2010年湖南省GDP指标量为:经查询知道湖南省2010年GDP指标量为15902.12亿元,超出95的预测区间,但还是相差不大。基本认为模型能够很好的表示该时间序列。预测图如下:4结论 本文使用时间序列分析的方法对湖南省国内生产总值的年度数据序列进行了随机性分析,并运用模型预测方法对我国的国内生产总值进行了小规模的预测。 通过模型识别、比较以及检验,最终选定ARIMA(2,2,0)模型:从该论证过程可以看出,在对经济指标做预测时,往往面对不平稳的时间序列模型,我们要
2、进行多阶差分之后,才能得到平稳的序列,建立理想的预测模型。参考文献1王燕.应用时间序列分析2徐国祥统计预测和决策(第二版)3赵蕾.ARIMA模型在福建省GDP预测中的应用4中国统计年鉴2010,附:所用数据湖南省历年GDP指标年份GDP人均GDP(元)GDP(亿元)占全国比重增幅(%)全国(本币)湖南本币美元比重(%)位次1978146.9987.294.231111.716.43812861701979178.01114.487.69.14193432211980191.72127.987.85.24633652441981209.68122.985.54923942311982232.52
3、122.83129.45284302271983257.43130.2810.99.25834702381984287.29123.4615.26955192241985349.95119.1513.512.08586262131986397.68115.178.98.19637032041987469.44126.1311.69.31,1128182201988584.07156.9211.38.21,3669992681989640.80170.204.13.61,5191,0742851990744.44155.643.84.01,6441,2282571991833.30156.557
4、.91,8931,3572551992986.98178.9314.211.12,3111,59528919931,244.71216.0214.012.42,9981,99734719941,650.02191.4413.110.64,0442,63030519952,132.13255.311010.35,0463,35940219962,540.13305.5210.012.15,8463,96347719972,849.27343.703.736,4204,42053319983,025.53365.453.668.56,7964,66756419993,214.54388.323.6
5、48.47,1594,93359620003,551.49428.983.619.07,8585,42565520013,831.90462.963.538.38,6226,12073920024,151.54501.583.44139,3986,73481420034,659.99563.003.309.610,5427,58991720045,641.94681.643.3710.112,3369,1651,10720056,596.10805.223.4712.214,18510,5621,28920067,688.67964.2812.712.816,50012,1391,522200
6、79,439.601,255.0215.020,16914,9591,989200811,555.001,663.7613.923,70818,1472,613200913,059.691,911.833.5813.725,57520,4282,990程序data new;input s;t=_n_;cards;1244.711650.022132.132540.132849.273025.533214.543551.493831.904151.544659.995641.946596.107688.679439.6011555.0013059.69;proc gplot data=new;p
7、lot s*t;symbol c=red v=circle i=jion;run;proc arima data=new;identify var=s nlag=8;difs=dif(s);数据plot s*t difs*t;symbol v=star c=black i=jion;identify var=difs nlag=8;difs=dif(dif(s);identify var=difs nlag=8 minic p=(0:5) q=(0:5);identify var=s(1,1) nlag=8;estimate p=(0,2) ;forecast lead=1 out=jj;data r; do t=1 to 33; output; end;data kk; merge r jj;proc gplot data=kk;plot s*t=1 forecast*t=2 l95*t=3 u95*t=3/overlay;symbol1 c=black i=none v=star;symbol2 c=red i=jion v=none;symbol3 c=green i=jion v=none l=32; (本资料素材和资料部分来自网络,仅供参考。请预览后才下载,期待您的好评与关注!)
copyright@ 2008-2022 冰豆网网站版权所有
经营许可证编号:鄂ICP备2022015515号-1