您好,欢迎访问三七文档
当前位置:首页 > 商业/管理/HR > 市场营销 > 美国1960年-1982年鸡肉需求模型分析
美国1960年-1982年鸡肉需求模型分析一、案例简介表1给出了影响美国鸡肉消费需求的变量数据,应变量是人均鸡肉消费量,解释变量包括人均真实可支配收入,鸡肉的价格,其他替代品的价格(猪肉和牛肉)。表1影响美国鸡肉消费需求的变量数据注:y=人均鸡肉消费(磅)X2=人均实际可支配收入(美元)X3=鸡肉实际零售价格(美分/磅)X4=猪肉实际零售价格(美分/磅)X5=牛肉实际零售价格(美分/磅)X6=鸡肉替代品的实际综合零售价格(美分/磅)obsYX2X3X4X5X6196027.8397.542.250.778.365.8196129.9413.338.15279.266.9196229.8439.240.35479.267.8196330.8459.739.555.379.269.6196431.2492.937.354.777.468.7196533.3528.638.163.780.273.6196635.6560.339.369.880.476.3196736.4624.637.865.983.977.2196836.7666.438.464.585.578.1196938.4717.840.17093.784.7197040.4768.238.673.2106.193.3197140.3843.339.867.8104.889.7197241.8911.639.779.1114100.7197340.4931.152.195.4124.1113.5197440.71021.548.994.2127.6115.3197540.11165.958.3123.5142.9136.7197642.71349.657.9129.9143.6139.2197744.11449.456.5117.6139.2132197846.71575.563.7130.9165.5132.1197950.61759.161.6129.8203.3154.4198050.11994.258.9128219.6174.9198151.72258.166.4141221.6180.8198252.92478.770.4168.2232.6189.4二、回归分析DependentVariable:YMethod:LeastSquaresDate:06/07/08Time:10:26Sample:19601982Includedobservations:23VariableCoefficientStd.Errort-StatisticProb.X20.0048890.0049620.9853700.3383X3-0.6518880.174400-3.7378890.0016X40.2432420.0895442.7164430.0147X50.1043180.0706441.4766740.1580X6-0.0711100.098381-0.7228050.4796C38.596914.2144889.1581500.0000R-squared0.944292Meandependentvar39.66957AdjustedR-squared0.927908S.D.dependentvar7.372950S.E.ofregression1.979635Akaikeinfocriterion4.423160Sumsquaredresid66.62224Schwarzcriterion4.719376Loglikelihood-44.86635F-statistic57.63303Durbin-Watsonstat1.100559Prob(F-statistic)0.000000图1回归分析结果由回归分析结果可知:①计算结果的标准形式:Y=38.59690942+0.004889344622*X2-0.6518875293*X3+0.2432418207*X4+0.1043176111*X5-0.07111034011*X6s=(0.004962)(0.174400)(0.08954)(0.070644)(0.098381)(4.214488)t=(0.985370)(-3.737889)(2.716443)(1.476674)(-0.722805)(9.158150)R-squared=0.944292F=57.63303DW=1.100559S.E.=1.979635②以b1、b2的置信区间为例:在a=5%的显著性水平上,查自由度为17的t分布表,的临界值ta/20.05(23-5-1)=1.7396,所以,置信度为95%的b1的置信区间为(b1^-ta/20.05(23-5-1)·s(b1^),b1^+ta/20.05(23-5-1)·s(b1^))=(0.004889-1.7396*0.174400,0.004889+1.7396*0.174400)=(-0.2978,0.3083)同理,置信度为95%的b1的置信区间为(b2^-ta/20.05(23-5-1)·s(b2^),b2^+ta/20.05(23-5-1)·s(b2^))=(-0.651888-1.7396*(0.08954),-0.651888+1.7396*(0.08954))=(-0.80765,-0.49612),其他类似。③模型的经济意义检验:回归系数估计值b1^=0.004889>0,说明人均实际可支配收入与鸡肉需求量正方向变化,当其他条件不变时,人均实际可支配收入上升1美元,对鸡肉的需求量增加0.004889磅;回归系数估计值b2^=-0.651888<0,说明鸡肉实际零售价格与鸡肉需求量反方向变化,当其他条件不变时,鸡肉实际零售价格上升1美元,对鸡肉的需求量减少0.651888磅;回归系数估计值b3^=0.243242>0,b4^=0.104318>0,说明猪肉和牛肉的需求量与消费者的人均可支配收入成正方向变化,当其他条件不变时,人均实际可支配收入上升1美元,对猪肉、牛肉的需求量分别增加0.243242磅、0.104318磅。④回归方程的标准误差的评价:S.E.=1.979635,说明回归方程与各观测点的平均误差为1.979635磅。⑤拟合优度检验:AdjustedR-squared=0.927908,说明回归方程即上述样本需求函数的解释能力为92.8%,回归方程的拟合优度较好。⑥回归模型的总体显著性检验:从全部因素的总体影响看,在a=5%的显著性水平上,F=57.63303>F0.05(5,17)=2.81,说明人均实际可支配收入、鸡肉实际零售价格对鸡肉的需求量的共同影响是显著的。从p值极小可以明显看出,只要显著性水平稍高,鸡肉零售价格和消费者实际可支配收入水平对鸡肉需求的共同影响是显著。⑦P值的显著性检验:从单个因素的影响上看,在a=5%的显著性水平上,t(b1^)=0.985370<ta/20.05(23-5-1)=1.7396,︳t(b2^)︱=︳-3.737889︱>ta/20.05(23-5-1)=1.7396,说明人均实际可支配收入对鸡肉需求的影响是不显著的,而鸡肉实际零售价格对鸡肉需求的影响是显著的;从图中可以看出人均实际可支配收入、鸡肉实际零售价格对应的P值分别为0.3383、0.0016,说明人均实际可支配收入对鸡肉需求的影响是不显著的,而鸡肉实际零售价格对鸡肉需求的影响是显著的。196027.828.8521390523-1.05213905231|.*|.|196129.931.9340084101-2.03400841014|*.|.|196229.831.0489742069-1.24897420686|.*|.|196330.831.8589315498-1.05893154979|.*|.|196431.233.1856928693-1.98569286929|*.|.|196533.334.9715574801-1.6715574801|*|.|196635.635.65692538-0.05692537995|.*.|196736.436.30161076490.098389235088|.*.|196836.735.87722317520.822776824816|.|*.|196938.436.74423286941.6557671306|.|*.|197040.439.42885041130.971149588668|.|*.|197140.337.82065365532.4793463447|.|.*|197241.841.14592550120.654074498787|.|*.|197340.437.26609955533.1339004447|.|.*|197440.739.73935924450.960640755455|.|*.|197540.141.5189213521-1.41892135214|.*|.|197642.744.1298431012-1.42984310116|.*|.|197744.142.59156480021.50843519977|.|*.|197846.744.48607930022.21392069978|.|.*|197950.648.84260589671.75739410334|.|*|198050.151.5569669577-1.45696695768|.*|.|198151.750.90933641920.790663580822|.|*.|198252.956.5324980471-3.6324980471|*.|.|图2拟合值和残差值三、多重共线性1、多重共线性检验(1)相关系数检验:在a=5%的显著性水平上,F=57.63303>F(5,17)=3.24,说明人均实际可支配收入、鸡肉实际零售价格对鸡肉的需求量的之间关系是显著的。CorrelationMatrix1YX2X3X4X5X6Y10.94717070.8399579450.9123918980620.935355440.937412988X20.94717075410.9316807840.9571311974710.9858775140.982757078X30.839957940.93168078410.9701116005180.9284688760.944528871X40.9123918980.9571311970.970111600510.9405665020.972964909X50.935355440.9858775140.92846887620.94056650226310.983348777X60.9374129880.9827570780.94452887160.9729649092310.983348771图3相关系数矩阵由图中可以看出,解释变量之间存在高度线性相关,和图1进行比较发现,尽管整体拟合度较好,但x的参数t值并不显著,表明解释变量之间存在多重共线性。(2)方差膨胀因子检验:VIF1=1/1-0.944292=15.3845,VIF2=1/1-0.83995794=5.8823,VIF3=1/1-0.912391898062=11.49425,VIF4=1/1-0.93535544=13.3354,VIF5=1/1-0.937412988=13.6982,方差膨胀因子大于10,因此,模型存在高度的多重共线性。2、多重共线性的修正(1)运用OLS方法逐一求y对各个解释变量的回归,结合经济意义和统计检验,选出拟合效果最好的一元线性回归方程,经分析,在5个一元回归模型中,人均鸡肉消费y对鸡肉零售价格x3的线性关系强,拟合优度好,回归结果如下:DependentVariable:YMethod:LeastSquaresDate:06/07/08Time:11:48Sample:19601982Includedobservations:23VariableCoefficient
本文标题:美国1960年-1982年鸡肉需求模型分析
链接地址:https://www.777doc.com/doc-1814131 .html