資源描述:
《計量經(jīng)濟(jì)學(xué)--多重共線性檢驗修正》由會員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在行業(yè)資料-天天文庫。
1、計量經(jīng)濟(jì)學(xué)實驗報告姓名:學(xué)號:班級:一、實驗?zāi)康谋緦嶒炑芯康氖菄鴥?nèi)旅游收入Y(億元)與國內(nèi)旅游人數(shù)X1(萬人次)、城鎮(zhèn)居民人均旅游支出X2(元)、農(nóng)村居民人均旅游支出X3(元)、公路里程X4(萬公里)和鐵路里程X5(萬公里)之間的關(guān)系,以便更好地了解和預(yù)測我國國內(nèi)旅游收入。表1是本體所用數(shù)據(jù),數(shù)據(jù)來源:《中國統(tǒng)計年鑒2004》。表1YX1X2X3X4X51023.552400414.754.9111.785.91375.76290046461.5115.75.971638.463900534.170.5118.586.492112.764400599.8145.7122.64
2、6.62391.269450607197127.856.642831.971900614.8249.5135.176.743175.574400678.6226.6140.276.873522.478400708.3212.7169.87.013878.487800739.7209.1176.527.193442.387000684.9200180.987.3二、模型假定先假定該模型是多遠(yuǎn)線性回歸模型:三、實驗步驟1.將數(shù)據(jù)導(dǎo)入數(shù)據(jù),得到表2。表22、在EViews命令框中直接鍵入“l(fā)sycx1x2x3x4x5”,按回車,即出現(xiàn)回歸結(jié)果表3。表3DependentVariab
3、le:YMethod:LeastSquaresDate:11/18/13Time:14:36Sample:110Includedobservations:10CoefficientStd.Errort-StatisticProb.??C-274.37731316.690-0.2083840.8451X10.0130880.0126921.0311720.3607X25.4381931.3803953.9395910.0170X33.2717730.9442153.4650730.0257X412.986244.1779293.1082960.0359X5-563.107732
4、1.2830-1.7526850.1545R-squared0.995406????Meandependentvar2539.200AdjustedR-squared0.989664????S.D.dependentvar985.0327S.E.ofregression100.1433????Akaikeinfocriterion12.33479Sumsquaredresid40114.74????Schwarzcriterion12.51634Loglikelihood-55.67396????Hannan-Quinncriter.12.13563F-statistic17
5、3.3525????Durbin-Watsonstat2.311565Prob(F-statistic)0.000092該模型的可絕系數(shù)高,F(xiàn)檢驗值是173.3525,擬合優(yōu)度好。但是,如果,的P值都大于0.05,不顯著,可見五個解釋變量之間存在多重共線性,模型需要進(jìn)行進(jìn)一步調(diào)整。3、計算解釋斌梁之間的相關(guān)系數(shù)矩陣,如表4。表4X1X2X3X4X5X110.9190.7520.9480.942X20.91910.8650.8590.963X30.7520.86510.6650.818X40.9480.8590.66510.898X50.9420.9630.8180.8981由
6、表可知。X1與X2、X4、X5的相關(guān)系數(shù)達(dá)到0.9以上,這充分表明各解釋變量間確實存在嚴(yán)重的多重共線性。4、五個解釋變量分別與被解釋變量進(jìn)行回歸分析,對結(jié)果整理得到表5。表5變量t值P值F值修正后的可決系數(shù)X18.66590.000075.09700.8917X213.15980.0000173.18020.9503X35.19670.000827.00530.7429X46.46750.000241.82800.8194X58.74870.000076.53970.8935表5是模型的檢驗,可見的擬合優(yōu)度最高,所以選該模型。5、進(jìn)行模型的檢驗,整理結(jié)果得出表6。表6變量t值
7、P值F值修正后的可決系數(shù)X24.28390.0036128.16690.9658X12.15120.0685X25.74790.000784.09780.9486X30.85780.4194X26.64460.0003155.79510.9718X42.65840.0325X22.90860.022778.18480.9449X50.46210.6580由表6可知,當(dāng)時,只有X2和X4這一組的P值都大于0.05,且該組的修正后的可決系數(shù)為最大值,擬合優(yōu)度最好,所以選擇模型是合理的。6、進(jìn)行模型的檢驗,整