时间序列分析Eviews

1、实验报告课程名称：时间序列分析院 系：班 级：姓 名：学 号：指导教师：设计时间：（一）时序图检验根据平稳时间序列均值、方差为常数的性质，平稳序列的时序图应该显示岀该序列始终在一个常数值附近随机波动, 而且波动的范围有界、无明显趋势及周期特征。view proc | object | properties print nbitie i freeze sample | ge nshut graph | stats ide nt图1:时序图（二）自相关图检验可以看出自相关系数始终在零周围波动，判定该序列为平稳吋间序列。t/i皀w | prou | object | properties | pri

2、nt | nzme | fri皀ne | sample | g亡nr | sh亡亡t| graph | stats | iderit |correlogram of x oarripre:iuxjincluded observations: 300autocorrelationpartial correlationacpacq-statprob厂i厂110.3310.33133.1430.00011c120.020-0.10133.2610.00011 11130.0390.01533.7320.000111'40.0090.03433.7590.00011|50.0410.06734

3、.2780.0001111160.0390.00334.7370.000|【i|17-0.040-0.06435.2430.00011'80.1010.06638.3880.000|i119-0.063-0.00939.6050.000111'110-0.0280.05039.8570.00011i|匚111-0.040-0.08340.3650.000|1 111120.0360.00740.7660.000ici応113-0.066-0.07042.1510.00011140.1260.11447.1890.000ici1115-0.0810.00949.2510.0001

4、|11160.0530.03750.1610.0001i11170.0040.02350.1670.0001111180.0060.01050.1800.0001111190.008-0.00150.1980.0001i11200.0010.00750.1990.000111'121>0.044-0.05150.8350.00011 1| 11220.0370.02551.2740.000匸1|匚123-0.100-0.09754.5420.00011240.1080.05958.3700.000匚1c125>0.155-0.10666.2380.000111260.099

5、0.03769.4560.000|11127-0.058-0.01670.5700.0001 i|128-0.033-0.06170.9390.000111【129-0.012-0.04870.9900.000ici匚130-0.080-0.12573.1350.0001'11310.0710.00974.8540.000|i'c132-0.065-0.07676.2860.000il1c1330.0730.10278.0880.000|1 1|1340.038-0.05978.5810.000i11135-0.0080.00178.6000.0001>1|360.044

6、0.04579.2650.000图2：序列的相关分析结果（三）平稳性检验一（单位根检验）view | proc | object | prop已rtiis print name freeze | sample | ginsheet | graph | stats identaugmented dickey-fuller unit root test on xnull hypothesis: x has a unit rootexogenous: constantlag length: 0 (automatic based on sic, maxlag=15)t-statisticprob?au

7、gmented dickey-fuller test statistic-24.299040.0000test critical values:1 % level 5% level 10% level-3.452066-2.870996-2.571880mackinnon (1996) one-sided p-values.augmented dickey-fuller test equation dependent variable: d(x) method: least squaresdate: 06/12/13 time: 20:10sample (adjusted): 2 300in

8、eluded observati ons: 299 after adjustme ntsvariablecoefficientstd. error t-statisticprob.x(-1)-1.3308780.054771-24.299040.0000c-0.0005950.000200-2.9792830.0031r-squared0.665331mean dependentvar3.34e-06adjusted r-squared0.664204s.d. dependent var0.005917s.e. of regression0.003429akaike info criterio

9、n-8.506493sum squared resid0.003492schwarz errteri on-8.481741log likelihood1273.721hannan-quinn criter.-8.496586f-statistic590.4435durbin-watson stat2.059758prob(f-statistic)0.000000图3：单位根检验结果由图3可见,检验t统计量的值为-24. 29904,显著性水平1%、5%、10%的临界值分别为-3. 452066> -2. 870996.-2. 571980, 可见t统计量的值小于各显著性水平的临界值，故

10、拒绝原假设，认为序列平稳。(四) 模型识别根据图2可见自相关系数与偏自相关系数的情况,我们分别选取ar、ma、arma (1,1)进行拟合,从中选优。vi皀w | proc| cfcjject| print| name | freeze | estimate forecaststats|re或ds|dependent variable: xmethod: least squaresdate: 06/11/13 time: 20:05sample (adjusted): 2 300included observations: 299 after adjustmentsconvergence ac

11、hieved after 3 iterationsvariablecoefficientstd. errort-statisticprob.c-0.0004470.000149-3.0021900.0029ar(1)0.3308780.054771-6.0411390.0000r-squared0.109433fvlean dependent var-0.000446adjusted r-squared0.106434sq dependent var0.003627s.e. of regression0.003429aka ike info criterion-8.506493sum squa

12、red resid0.003492schwarz criterion-8.481741log likelihood1273 721hannan-quinn criter.-8.496586f-statistic36.49536durbin-watson stat2.059758pro b(f-statistic)0.000000inverted ar roots.33图4： ar (1)参数估计结果view | proc | object |print | name | freeze |estimate | forecast |stats | resids |dependent variabl

13、e: xmethod: least squaresdate: 06/11/13 time: 20:09sample: 1 300included observations: 300convergence achieved after 7 iterations ma backcast: 0variablecoefficientstd. errort-statisticprob.c-0.0004420.000130-3.4017700.0008ma(1)-0.3412500.054740-6.2340450.0000r-squared adjusted r-squared s.e. of regr

14、ession sum squared resid log likelihood f-statisticp rob(f-stati sti c)0.1146950.1117240.0034130.0034711279.36638.607160.000000mean dependent var sq dependent var aka ike info criterion schwarz criterion hannan-quinn criter. durbin-watson stat-0.0004450.003621 -8.515772 -8.4-91080 -8.5058902.036594i

15、nverted ma roots.34图5： ma(1)参数估计结果vi已w | obj已ut | printj zem已 | 尸厂已已n亡 | estimbt已 | foeubst| stats | r已sids |dependent variable: xmethod: least squaresdate: 06/h/13 time: 21:34sample (adjlisted): 2 300included observations: 299 after adjustmentsconvergence achieved after d 4 iterationsma backcast: 1

16、variablecoefficientstd. errort-statisticprob.c-0.0004440.000134-3.3048000.0011ar(1)0.1248770.167309-0.7463860.4560ma(1)-0.2363230.164632-1.4354650.1522r-squared0.11691 1fvlean dependent var0.000446adjusted r-squaredo.xo944s.d dependent var0.003627s.e of regression0.003420aka ike info criterion-8.508

17、236sum squared resid0.003463schwarz criterion-8.471108log likelihood127-4.981hannan-quinn criter.-8.493376f-statistic19.59345durbin-watson stat1.999662p ro b(f-stati sti c>0.000000inverted ar roots-.12inverted ma roots.24图6： arma(1,1)参数估计结果注：(由p值可知此模型未能很好的拟合这组数据，故舍弃。) 从aic方法看，以上三个模型的拟合中，模型am(l)较适

18、合。 模型为 x=-0.000442-0.34l250模型的单位根如下：单位根在单位圆内，所以可逆图7：单位根结果（五）模型检验1、参数的显著性检验根据图5中的p值检验参数的显著性，均小于0.05时，拒绝原假设，即参数显著不为0。2、检验 £ t是否为白噪声序列（1）检验均值是否为零，由于p值=0.9999,所以接受原假设，即期望值为0。 series: resid workfile: untitled:untitled| o | 回 |衣3view | procobjectpropertiesprint |name | freeze |sample | genr sheet | graph | stats | iden11hypothesis testing for resid date: 06/12/13 time: 20:44 sample: 1 300included observations: 300test of hypothesis: fvlean = 0.000000sample rvlean = -1.74e-08sample std. dev.二 cl003621methodvalueprobabilitystatistic-8