eviews6.0實(shí)際案例操作

1、Eviews6.0實(shí)際案例操作安裝好eviews6.0案例：19782000年我國財(cái)政收入Y和GDP的關(guān)系19781132.26019783624.10019791146.38019794038.20019801159.93019804517.800119811175.7901981486240019821212.33019825294.70019831366.95019835934.50019841642.86019847171.00019852004.82019858964.40019862122.010198610202.2019872199.350198711962.501988235

2、7.240198814928.30|19892664.900198916909.2019902937.100199018547.9019913U9.480199121617.80J19923483.370199226638.1019934348.950199334634.4019945218.100199446759.4019956242.200199558478.1019967407.9901996678846019978651.140199774462.6019989875.950199878345.20199911444.08199982067.50200013395.232000894

3、03.60YGDP回歸分析操作步驟、FileNewworkfile二、輸入樣本的startdate_enddate開始結(jié)束時(shí)間WF:給回歸方程取一個(gè)名字三、quick一emptygroup輸入數(shù)據(jù)2/93/9Range:19782000-23obsSample:19782000一23obsddp引cgdreyEEIMIS親斤増了y和gdp（開始顯示是serOl通過右擊rename修改名？）四、回歸分析ObjectnewobjectOK輸入“ycgdp"確定得到卜圖結(jié)果:#/9Equation:UHTITLEDErk珀論Y-GDPS=nnt.口回岡亡7Poc：Eobjict1(Pint

4、)(N”亡(F亡亡z亡(Estimate)(Ft(StatsResidsDependentVariable:YMethod:LeastSquaresDate:04/18/12Time:18:45Sample:19782000Includedobservations:23CoefficientStd.Errort-St3tisticProb.C556.6477220.89432.5199730.0199GDP0.1198070.00527322.722980.0000R-squared0.960918Meandependentvar4188.627AdjustedR-squared0.95905

5、7S.D.dependentvar3613.700S.E.ofregression731.2086Akaikeinfocriterion16.11022Sumsquaredresid11227988Schwarzcriterion16.20895Loglikelihood-183.2675Hannan-Quinncriter.16.13505F-statistic516.3338Durbin-Watsonstat0.347372Prob(F-statistic)0.000000X=556.65+0.1198GDI>（2.52）（22.72）R2=0.9609Dependentvariab

6、le：被解釋變戢Method:LeastSquares最小二乘法Includedobservations:A觀察對象個(gè)數(shù)Coefficient：系數(shù)Std.Error:標(biāo)準(zhǔn)謀T-statistic：t統(tǒng)計(jì)呈（1）T-Statistic=平均值/標(biāo)準(zhǔn)誤，（2）標(biāo)準(zhǔn)誤二標(biāo)準(zhǔn)差/樣本呈的開平方。T-Statistic中的平均值指的是平均至值之類的概念，有單樣本t檢驗(yàn)和雙樣木t檢驗(yàn)之分，具休如下。（A）單樣本t檢驗(yàn)：Statistic二（樣本平均值靶值）/標(biāo)準(zhǔn)誤靶值就是你想要與你的樣本均值相比較的數(shù)值。在單樣木t檢驗(yàn)中，T-statistic的絕對值越大，則樣木平均值偏離靶值越遠(yuǎn)，也就是樣木平均值與靶

7、值有顯著壟異的概率越大。（B）雙樣木t檢驗(yàn)：T-Statistic=（樣木1的平均值-樣本2的平均值）/標(biāo)準(zhǔn)誤T-statistic用丁檢驗(yàn)左異是否有顯著性，其數(shù)值大小不存在好壞之分，具休意義如下。在雙樣木t檢驗(yàn)中，T-statistic的絕對值越大，則兩組樣木平均值的基距越大，也就是兩組樣本平均值有顯著壟異的概率越大。）Prob：p值R_squared:擬合優(yōu)度AdjustedR-squared：調(diào)整后的擬合優(yōu)度S.Eofregression：擾動項(xiàng)標(biāo)準(zhǔn)基sumsquaredresid殘?bào)咂椒亢虴SSloglikelihood:對數(shù)似然函數(shù)值F-statistic：F統(tǒng)計(jì)呈（F檢驗(yàn)統(tǒng)計(jì)嵐是檢

8、驗(yàn)整個(gè)方程的顯著性的，即解耗平方和除以殘墊平方和，另外對丁方蔓齊性的檢驗(yàn)過程也采用的是F檢驗(yàn)）Prob（F-statistic）:F統(tǒng)計(jì)雖:的概率（F統(tǒng)計(jì)雖的雙尾或者單尾概率值）建立圖像Quickgraph輸入“gdpy”選擇scatter得到散點(diǎn)圖回歸方程的檢驗(yàn)在|'1山度為5%的顯著性水平下，|'|山度為23-2=21的t分布的臨界值為2.08,而截距項(xiàng)的t統(tǒng)計(jì)噬值為2.52>2.08,斜率的t統(tǒng)計(jì)戢為22.72>2.08個(gè)此,兩參數(shù)在統(tǒng)計(jì)雖上是顯著的。列外，樣本可決系數(shù)R2=0.9609表明，財(cái)政收入96%的變化可以山國內(nèi)生產(chǎn)總值的變化來解釋，回歸直線的擬合程

9、度很好。預(yù)測根據(jù)回歸模型乂=556.65+O.1198GDE,當(dāng)2001年GDP值為105709億元時(shí)，財(cái)政收入預(yù)測值：13220.59±2.08x425.75進(jìn)行單值的區(qū)間預(yù)測7/9MM冊mWIM關(guān)加i網(wǎng)寸(Pr°c(0bj亡ct)(Print)(Nom亡亡亡z亡Estimat亡(Fo亡castHstatsResids8/9Dependent仙iableYMethod:LeastSquaresDate:04/18/12Time:20:16Sample(adjusted):19782000Ineludedobservations:23afteradjustments點(diǎn)擊上面

10、回歸分析的"forecast"CoefficientSid.Errort-StatisticProb.c556.6477220.89432.5199730.0199GDP0.1198070.00527322.722980.0000R-squared0.960918hleandependent4188.627AdjustedR-squared0.959057S.D.dependentvar3613.700S.E.ofregression731.2086Akaikeinfocriterion16.11022Sumsquaredresid11227988Schwarzcriter

11、ion16.20895Loglikelihood-183.2675Hnnan-Quinncriter.16.13505F-statistic516.3338Durbin-Watsonstat0.347372Prob(F-statistic)0.000000在forecastname輸入：yf在s.e輸入：yfse得到如下結(jié)果：|Y泡口|PnoJ|Objwet|Pint|SawE&t3il£十卜Show(F57|Sto佗|“1住|!7射叩1孑DisplayFilter:Range:19782001-24obsSample:19782001-24obsc0gdp0resid09/

12、9Syf之yfse#/9#/9<?UntitledXNewRqgw/打開yf序列査看2001年預(yù)測值:Lastupdated:04/18/12-20:19Modified:19782001/fitff=actual)yf1978990.841619791040.45419801097.91319811139.19919821190.99219831267.64519841415.78619851630.64919861778.94719871989.84319882345.16819892582.49519902778.82319913146.62019923748.0891993470

13、6.10419946158.76919957562.75519968689.72319979477.81619989942.980199910388.94200011267.86200113221.372001年的財(cái)政收入為13221.37(與計(jì)算所得結(jié)果13220.59不同的療因是，按照公式計(jì)算時(shí)我們用的4=0.1198,而實(shí)際上是0.119807)標(biāo)準(zhǔn)廷為頭=846.13846.12711990749.50771991748.3417199274728251994751.95071995761.55211996772.75511997782.36411998788

14、.69681999795.21562000809.30392001846.1271代入公式預(yù)測：&OO1-t0025x每001»XoOl+0.025X)結(jié)果為(11460.59,14980.54)最后預(yù)測財(cái)政收入均值的咒信區(qū)間，預(yù)測的均值的標(biāo)準(zhǔn)墊為：干擾項(xiàng)的標(biāo)準(zhǔn)(S.E.ofregression)為:731.2086EquationsUHIITLEDTorkfile:T-GDP:sUnt.-.口回區(qū))vie：八卩0匚)(0可亡ct)pint)(N6m亡f亡亡z亡(Estimate)(F|statsResiclsDependentVai2ble:YMethod:LeastSqu

15、aresDate:04/18/12Time:18:45Sample:19782000Includedobservations:23CoefficientStd.Errort-SiatisticProb.C556.6477220.89432.5199730.0199GDP0.1198070.00527322.722980.0000R-squared0.960918Meandependentvar4188.627AdjustedR-squared0.959057S.D.dependentvar3613.700S.E.ofregression731.2086Akaikeinfocriterion16.11022Sumsquaredresid11227988Schwarzcriterion16.20895Loglikelihood-183