课程论文分析题目:中国粮食产量影响因素分析报告学院:经济学院专业:经济学一班学号:201301010111姓名:刘明华指导教师:秦翊2015年6月6日ShanxiuniversityofFinanceandEconomics修德立信博学求真一、问题的提出改革开放以来,中国经济迅速发展,人口增长迅猛,对粮食的需求日益增加。粮食产量无疑成了影响中国经济发展的重大因素。同时,粮食的产量直接关系到农业劳动力的生活水平,因此,“三农”问题成为中国经济研究的热点问题,提高粮食产量,关注农村居民收入迫在眉睫。为此,本文将就粮食产量影响因素进行分析,希望从中发现一些对粮食产量关键作用的因素。二、模型设定及假设通过对影响粮食产量的主要因素的分析,把影响农民收入的因素主要归结与以下几个方面:农业化肥施用量X1,粮食播种面积X2,成灾面积X3,农业机械总动力X4,农业劳动力X5。为此设定了如下形式的计量经济学模型:01122334455tttttttY其中Yt为第t年粮食产量(万吨),X1为农业化肥使用量(万公顷),X2粮食播种面积(千公顷),X3为成灾面积(公顷),X4为农业机械总动力(万千瓦),X5为农业劳动力(万人)。三、数据的收集通过查找中国统计年鉴,我们得到如下的统计资料:年份粮食产量(万吨)农业化肥施用量(万公斤)粮食播种面积(千公顷)成灾面积(公顷)农业机械总动力(万千瓦)农业劳动力(万人)19853872816601140471620918022311511986407311740112884152641949730868198737911177610884522705209133113019883915119311109332365622950312541989402081999111268203932483631663199039408214211012323945265753224919914075523571122052444928067332251992446242590113466178192870838914199343529280611231427814293893909819944426429301105602589530308386991995456493152110509231333181737680199644510331810954431383338023662819974666235941100602226736118355301998504543828112548212333854734820199949417398111291230309420163484020005123040841137872518145208351772001508394124113161267314899635768中国粮食生产与相关投入资料四、模型的估计与调整主要需要检验的有:一、多重共线性检验。二、异方差性检验。三、自相关性检验。利用Eviews5.0作OLS估计的结果为:DependentVariable:YMethod:LeastSquaresDate:06/12/15Time:13:36Sample:19852009Includedobservations:25VariableCoefficientStd.Errort-StatisticProb.X15.9945110.6097139.8316850.0000X20.5367010.0578589.2762450.0000X3-0.1358730.029720-4.5717320.0002X4-0.0908220.042053-2.1596960.0438X5-0.0073900.070511-0.1048140.9176C-26695.087507.527-3.5557750.0021R-squared0.980829Meandependentvar44945.64AdjustedR-squared0.975783S.D.dependentvar4150.729S.E.ofregression645.9230Akaikeinfocriterion15.98480Sumsquaredresid7927113.Schwarzcriterion16.27733Loglikelihood-193.8100Hannan-Quinncriter.16.06594F-statistic194.4114Durbin-Watsonstat1.715679Prob(F-statistic)0.000000Y=-26695.08+5.994511X1+0.536701X2-0.135873X3-0.090822X4-0.007390X5(7507.527)(0.609713)(0.057858)(0.029720)(0.042053)(0.070511)T=(-3.555775)(9.831685)(9.276245)(-4.571732)(-2.159696)(-0.104814)R-Squared=0.980829df=192002462184146108463343745257436043200345264425410608031793551723651320044570643391038912731957930368702005430704412994103251660387365462006469474637101606162976402835269200748402476610427819966683983397020084980449281049582463272522325612009501605108105638250647659031444由此可见:可决系数R-Squared=0.980829,表明模型在整体的拟和非常好。系数显著性检验:对于C、X1、X2、X3、X4的系数,t的统计量的绝对值都通过了检验,而X5的系数的t统计量为-0.104814,在df=19、α=0.05的情况下,X5的系数不能通过检验。根据经验判断,无法通过第一步检验的原因很可能是解释变量之间存在多重共线性。1、修正多重共线性我们对X1X2X3X4X5进行多重共线性检验,得到:相关系数表X1X2X3X4X5X11.000000-0.6165660.4006440.9527460.314885X2-0.6165661.000000-0.238039-0.741538-0.060970X30.400644-0.2380391.0000000.3100960.409704X40.952746-0.7415380.3100961.0000000.128834X50.314885-0.0609700.4097040.1288341.000000可以发现X1X2X3X4X5之间存在高度的线性相关关系。运用逐步回归法进行修正:一元回归估计结果变量X1X2X3X4X5参数估计值3.158761-0.144290.1827150.1652190.553797T统计量7.716525-0.682971.1265644.7750661.799071r^20.7213630.0198770.0522950.1652190.123364其中,加入X1的r^2最大,以X1为基础,顺次加入其他变量逐步回归。结果如下。加入新变量的回归结果(一)加入变量X2X3X4X5参数估计值0.631835-0.10622-0.262970.146656T统计量11.07516-1.11232-3.972170.79565r^20.9576240.7361990.8377370.729157其中,加入X2的r^2最大,以X1,X2为基础,顺次加入其他变量逐步回归。结果如下。加入新变量的回归结果(二)加入变量X3X4X5参数估计-0.11151-0.036810.002836值T统计量-3.63213-0.826050.037402r^20.9739740.9589580.957627其中,加入X3的r^2最大,以X1,X2,X3为基础,顺次加入其他变量逐步回归。加入新变量的回归结果(三)加入变量X4X5参数估计值-0.088210.082863t值-2.671131.34134r^20.9808170.082863显然可见,加入X5时,参数的检验值不显著,说明主要是因为X5引起了多重共线性。修正多重共线性以后的回归结果为:DependentVariable:YMethod:LeastSquaresDate:06/12/15Time:14:07Sample:19852009Includedobservations:25VariableCoefficientStd.Errort-StatisticProb.X15.9545330.46376912.839430.0000X20.5385190.05381610.006730.0000X3-0.1363930.028570-4.7739860.0001X40.0882100.033023-2.6711340.0147C-27110.396217.065-4.3606410.0003R-squared0.980817Meandependentvar44945.64AdjustedR-squared0.976981S.D.dependentvar4150.729S.E.ofregression629.7498Akaikeinfocriterion15.90538Sumsquaredresid7931696.Schwarzcriterion16.14915Loglikelihood-193.8172Durbin-Watsonstat1.706044F-statistic255.6537Prob(F-statistic)0.000000Y=-27110.39+5.954533X1+0.538519X2-0.136393X3+0.088210X4(6217.065)(0.463769)(0.053816)(0.028570)(0.033023)T=(-4.360641)(12.83943)(10.00673)(-4.571732)(-2.671134)R-Squared=0.980817AdjustedR-squared=0.976981F-statistic=255.65372、自相关检验-2,000-1,500-1,000-50005001,00036,00040,00044,00048,00052,000868890929496980002040608ResidualActualFittedDW检验:由表的DW=1.706044,在显著性水平=0.05下,查DW表,n=25,k=4,得到dl=1.038,dv=1.767,由于DW=1.706044,介于DL和DU之间,所以根据判定定理无法通过DW检验其自相关是否存在。3、异方差检验WhiteHeteroskedasticityTest:F-statistic1.292972Probability0.347081Obs*R-squared16.10371Probability0.307082TestEquation:DependentVariable:RESID^2Method:LeastSquaresDate:06/12/15Time:14:44Sample:19852009Includedobservations:25VariableCoefficientStd.Errort-StatisticProb.C-2.30E+082.75E+08-0.8375160.4219X149179.6741752.131.1778960.2661X1^2-0.9958862.105686-0.4729510.6464X1*X2-0.4443690.387229-1.1475610.2779X1*X3-0.