影响我国农业总产值因素的实证分析..docx

影响我国农业总产值因素的实证分析一、模型设定为了分析我国农业总产值及其影响因素的关系，选择“农业总产值”（单位：亿元）为被解释变量Y；选择“农村居民家庭生产型投资”（单位：元/每户）为解释变量X1；选择“农作物总播种面积”（单位：千公顷）为解释变量X2；选择“农业从业人员数”（单位：万人）为解释变量X3。实验数据来自国家统计局公布的数据为样本：我国农业总产值及其影响因素年份农业总产值Y（单位：亿元）农村居民家庭生产型投资X1（元/每户）农作物总播种面积X2（千公顷）农业从业人员数X3（万人5811269333225.0019904954.3012582738914.0019915146.4014018039098.0019925588.0016431038699.0019936605.1019507037680.0019949169.1923476036628.00199511884.6027743035530.00199613539.7536056034820.00199713852.5038962034840.00199814241.8839707035177.00199914106.2240458135768.00200013873.6046768536042.50200114462.8048838636398.50200214931.5452215136640.00200314870.1055869636204.40200418138.3659565534829.80200519613.3771557333441.90200621522.2876470031940.60200724658.1083899330731.00200828044.1590547029923.30200930777.5099705528890.50201036941.11107068127930.50201141988.64160872226594.00201246940.46169746725773.00201351497.37182009324171.00根据以上数据，利用Eviews生成Y、X1、X2、X3的线性图：由线形图可以看出，我国农业总产值及各影响因素的差异明显，其变动的方向基本相同，相互可能具有一定的相关性。将模型设定为线性回归模型形式：二、估计参数利用Eviews估计模型参数，回归结果如下：Dependent Variable: YMethod: Least SquaresDate: 05/30/15 Time: 20:54Sample: 1989 2013Included observations: 25VariableCoefficientStd. Errort-StatisticProb.??C-29869.2020836.69-1.4334910.1664X11.7098020.2323387.3591110.0000X20.4095780.1409762.9053050.0085X3-0.7415040.197329-3.7577010.0012R-squared0.989559????Mean dependent var19257.92Adjusted R-squared0.988067????S.D. dependent var13261.48S.E. of regression1448.648????Akaike info criterion17.54030Sum squared resi???Schwarz criterion17.73532Log likelihood-215.2537????Hannan-Quinn criter.17.59439F-statistic663.4221????Durbin-Watson stat1.889115Prob(F-statistic)0.000000根据上表数据，模型估计的初步结果为：（20837.69）（0.2323）（0.1410）（0.1973） t= （-1.4335）（7.3591）（2.9053）