时间序列实验报告3

  • 格式:pdf
  • 大小:223.37 KB
  • 文档页数:3

下载文档原格式

  / 3
  1. 1、下载文档前请自行甄别文档内容的完整性,平台不提供额外的编辑、内容补充、找答案等附加服务。
  2. 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
  3. 3、如文档侵犯您的权益,请联系客服反馈,我们会尽快为您处理(人工客服工作时间:9:00-18:30)。

时间序列分析实验报告

Problem1:Estimate ARMA-ARCH model for financial series in arch序列.xls

◆create new integer-data workfile named arch1,import the data series named y

◆Estimate AR model by correlogram--eq01

◆Diagnostic checks:test whether there is serial correlation in the residuals by

Q-statistics and LM–test.

◆Test heteroskedasticity,give your reason briefly

◆Establish AR-ARCH(q)model for possible order q,and select the best one as your

final model(note:parameters,whether there is remaining ARCH effect in standardized

residuals and information criterions should be considered)

◆Write out the mean equation and variance equation---eq02

Problem2:Estimate ARMA-TGARCH(EGARCH)model for financial series in杠杆数据.xls

◆create a new integer-data workfile named杠杆;import data series named y

◆Estimate ARMA model by correlogram----eq01—

(sometimes the significance of coefficient can be omitted temporarily)

1.Diagnostic checks:test whether there is serial correlation in the residuals by

Q-statistics and LM–test.

2.Test heteroskedasticity(null hypothesis(H0):there is no ARCH in the residuals)

◆Correlogram of squared residuals----Q-statistic

◆ARCH-LM test:(In the Lag Specification dialog box you should specify

the lag order)

✧Establish ARMA-TGARCH--eq02,check whether there is leverage effect,give

your reason:

◆diagnostic checking on standardized residuals of eq02.

◆Write out mean equation and variance equation of eq02

◆You can try to establish ARMA-EGARCH model,check whether there is leverage effect and

give your reason:---eq03

◆diagnostic checking on standardized residuals of eq03.

◆Write out mean equation and variance equation---eq03

实验报告结果

Yt=0.477*Yt-1-0.208*Yt-2+Ut

(1-0.447*L+0.208*L^2)Yt=Ut

which is calculated with12correlation coefficients(and10

degrees of freedom)is Q(12)=12.101.Since p value(=0.278)is larger than0.05,there is

no serial correlation in the residuals under the5percent level

Establish an AR(1)-ARCH(1)model

Yt=0.439Yt-1+ξt

Ht^2=1.133+0.980ξt-1^2

the corresponding p value is0.318.This implies that the squared standardized residuals are not auto correlated.(12)11.525Q

(3.2)In LM test,the value of the test statistic is LM(2)=0.041,and the co80rresponding p value is 0.980.This implies that there exists no ARCH effect in the standardized residuals.

Ϭ^2=1.113/(1-0.980)=56.650

通过指定LM检验滞后的阶数为2,发现残差中不存在自相关性,截图如下:

(2.2)通过对残差平方的LBQ检验发现,残差平方中存在自相关性,截图如下:

(2.3)通过在ARCH-LM检验中指定滞后的阶数为2发现,条件异方差性存在