经济分析工具综合训练(第一次作业)

经济分析工具综合训练（第一次作业）结合原始数据和中国统计局2000年至2012年公布的统计数据后得出总数据表如下：粮食生产模型样本观测值数据（1983年~2012年）年份（年）粮食总产量农用化肥施用量粮食播种面积成灾面积农业机械劳动力农业劳力（万人）X5（万吨）Y（万公斤）X1（千公顷）X2（千公顷）X3（万千瓦）X4198338727.51659.81140471620918022.131645.1198440730.51739.81128841560719497.231685198537910.81775.81088452270520912.530351.5198639151.21930.6110933236562295030467.9198740473.31999.31112682039324836308701988394082141.5110123239452657531455.71989407552357.1112205224492806732440.51990446242590.31134661781928707.733336.41991435292805.11123142781429388.634186.3199244265.82930.21105602585930308.434037199345648.83151.91105092313331816.633258.2199444510.13317.91085443138333802.532690.3199546661.83593.71100602226736118.132334.5199650453.53827.91125482123338546.932260.4199749417.13980.71129123030942015.632434.9199851229.54083.71137872518145207.732626.4199950838.64124.31131612673148996.132911.8200046217.54146.41084633437452573.632797.5200145263.74253.81060803179355172.132451.0200245705.84339.41038912716057929.931990.6200343069.54411.6994103251660386.531259.6200446947.04636.61016061629764027.930596.0200548402.24766.21042781996668397.929975.5200649804.24927.71049582463272522.129418.4200750160.35107.81056382506476589.628640.7200852870.95239.01067932228382190.428363.6200953082.15404.41089862123487496.128065.3201054647.75561.71098761853892780.527694.8201157120.95704.21105731241197734.727355.4201258958.05838.911120511470102559.027032.3一、利用表中数据，建立回归模型为避免多重共线性，要对部分解释变量进行替换。把粮食产量X2替换为有效种植面积X2-X3,把农业机械劳动力X4替换为单位面积农业机械劳动力X4/(X2-X3)，把农业化肥使用量X1替换为单位面积化肥使用量X1/(X2-X3)。：二、输出结果分析(一)、第一次输出结果分析第一次输出结果DependentVariable:LOG(Y)Method:LeastSquaresDate:03/11/15Time:20:36Sample:19832012Includedobservations:30VariableCoefficientStd.Errort-StatisticProb.C4.2024452.3332221.8011340.0838LOG(X2-X3)0.8454300.06492713.021260.0000LOG(X5)-0.1610180.171684-0.9378770.3573LOG(X1/(X2-X3))0.4653670.0754136.1709000.0000LOG(X4/(X2-X3))-0.1385540.068445-2.0242960.0537R-squared0.966189Meandependentvar10.74448AdjustedR-squared0.960779S.D.dependentvar0.117795S.E.ofregression0.023328Akaikeinfocriterion-4.527275Sumsquaredresid0.013605Schwarzcriterion-4.293742Loglikelihood72.90912Hannan-Quinncriter.-4.452566F-statistic178.6016Durbin-Watsonstat1.288652Prob(F-statistic)0.0000001、经济意义检验αj（j=1,…，5）均介于0~1之间，作为产出弹性，这是符合经济意义的。但X5和X4/(X2-X3)的预估系数的符号小于0，与事实相悖，考虑剔除这两个变量。2、统计检验拟合优度检验：可决系数R2=0.966，修正后可决系数为0.960，说明样本与预期线性模型拟合的很好。F检验：F0.05（4，25）=2.76178.6，解释变量对被解释变量有显著影响t检验：t0.25（25）=0.684，各解释变量的t检验值分别为1.801134、13.02126、-0.937877、6.170900、-2.024296。结果为全部显著。但t分布的情况是X5、X4/(X2-X3)的pro值为0.3573、0.05370.05，所以考虑剔除X5、X4/(X2-X3)。D.W.序列无关范围为1.65~2.35，显然存在序列相关(二)、第二次输出结果分析第二次输出结果分析（剔除X5）DependentVariable:LOG(Y)Method:LeastSquaresDate:03/11/15Time:21:03Sample:19832012Includedobservations:30VariableCoefficientStd.Errort-StatisticProb.C2.0905110.6094903.4299320.0020LOG(X2-X3)0.8720370.05826614.966530.0000LOG(X1/(X2-X3))0.4086330.0449279.0954370.0000LOG(X4/(X2-X3))-0.0825450.033366-2.4739380.0202R-squared0.964999Meandependentvar10.74448AdjustedR-squared0.960961S.D.dependentvar0.117795S.E.ofregression0.023274Akaikeinfocriterion-4.559362Sumsquaredresid0.014084Schwarzcriterion-4.372535Loglikelihood72.39043Hannan-Quinncriter.-4.499594F-statistic238.9487Durbin-Watsonstat1.292118Prob(F-statistic)0.000000R2=0.964，说明样本与预期线性模型拟合的较好。F0.05（3，26）=2.98238.9，解释变量对被解释变量有显著影响T检验中，所有变量均能通过但X4/(X2-X3)不能通过经济意义检验，考虑剔除X4/(X2-X3)。D.W.序列无关范围为1.57~2.43，显然存在序列相关(三)、第三次输出结果分析第三次输出结果分析（剔除X4/(X2-X3)）DependentVariable:LOG(Y)Method:LeastSquaresDate:03/11/15Time:21:25Sample:19832012Includedobservations:30VariableCoefficientStd.Errort-StatisticProb.C0.3695921.4423380.2562450.7998LOG(X2-X3)0.8726470.06719712.986360.0000LOG(X5)0.1422090.0887441.6024720.1211LOG(X1/(X2-X3))0.3154400.01502620.993110.0000R-squared0.960647Meandependentvar10.74448AdjustedR-squared0.956106S.D.dependentvar0.117795S.E.ofregression0.024679Akaikeinfocriterion-4.442155Sumsquaredresid0.015835Schwarzcriterion-4.255329Loglikelihood70.63233Hannan-Quinncriter.-4.382388F-statistic211.5629Durbin-Watsonstat1.153288Prob(F-statistic)0.000000R2=0.964，说明样本与预期线性模型拟合的较好F0.05（3，26）=2.98211.56，解释变量对被解释变量有显著影响三个变量均能通过经济意义检验t分布的情况是X5的pro值为0.12110.05，所以考虑剔除X5D.W.序列无关范围为1.57~2.43，显然存在序列相关(四)、第四次输出结果分析第四次输出结果分析（剔除X5、X4/(X2-X3)）DependentVariable:LOG(Y)Method:LeastSquaresDate:03/11/15Time:21:30Sample:19832012Includedobservations:30VariableCoefficientStd.Errort-StatisticProb.C2.4506420.6455373.7962860.0008LOG(X2-X3)0.8149190.05834913.966330.0000LOG(X1/(X2-X3))0.3011170.01242424.236490.0000R-squared0.956760Meandependentvar10.74448AdjustedR-squared0.953557S.D.dependentvar0.117795S.E.ofregression0.025386Akaikeinfocriterion-4.414634Sumsquaredresid0.017399Schwarzcriterion-4.274515Loglikelihood69.21952Hannan-Quinncriter.-4.369809F-statistic298.7138Durbin-Watsonstat0.933877Prob(F-statistic)0.000000两个变量都能够通过经济意义检验R2=0.956760，说明样本与预期线性模型拟合的较好F0.05（3，26）=2.98298.7，解释变量对被解释变量有显著影响t检验，所有变量均通过，即对被解释变量的影响均为显著D.W.序列无关范围为1.49~2.51，显然存在序列相关三、粮食生产模型结果模拟（不考虑序列相关）粮食生产模型模拟结果四、预测与分析