文档介绍：该【GARCH模型建模 】是由【guoxiachuanyue015】上传分享，文档一共【15】页，该文档可以免费在线阅读，需要了解更多关于【GARCH模型建模 】的内容，可以使用淘豆网的站内搜索功能，选择自己适合的文档，以下文字是截取该文章内的部分文字，如需要获得完整电子版，请下载此文档到您的设备，方便您编辑和打印。GARCH模型建模
以1995年1月至2000年8月日元兑美元汇率值序列ARCH建模分析(共1427个值)日元兑美元汇率值序列JPY
做单位根检验
■2
■2
Droc|objec:|Properties][Print
Name[FreezeSampleGenr
Spreadsheet
Graph...
DescripiiveStatistics&Tests►
One-WayTabulation...
Cnrr&logrann...
Lang-'unVarianoe...
UnitRootTest...
VarianceRatioTest..
BDSIndependenceTest..
Label
1
in
***@Series:Worlcfile:氏KH匚ASE12::Case12\
GARCH模型建模
GARCH模型建模
NullHypothesis:JPYhasaunitroot
Exogenous:Constant
LagLength:0(Automatic-basedonSIC,maxlag=23)
t-StatisticProb.*
AuqmentedDickey-Fulle-rteststatistic
-
Testcriticalvalues:
1%le-vel
-
S%level
-
10%level
-
^MacKinnon(1996)one-sidedp-values.
GARCH模型建模
GARCH模型建模
QuickOptionsAdd-insWindow
Help
Sample...
GenerateSeries...
Show...
E12
Graph...
)
-□X
Genr
kiNrlaRam
Serie5Statistics
►
GroupStatistics
EstimateEquation...
EstimateVAR,.,
EmptyGroup(EditSeri
les:.
/poth
nous:T;onslant
^ngth:D(Automatic-basedonSIC,maxis
HistogramandStats
UnitRootTest...
Variance-RatioTest...
ExponentialSmoothing...
Hodrick-PrescottFilter...
^Statistic
Prob.*
GARCH模型建模
GARCH模型建模
GARCH模型建模
GARCH模型建模
Aulcic;Drrelaban
PartialCorrelaiion
AC
RAC
Q-Siat
Prob
i
i
1
Q039
22127
I
1
I
2
i
L
1
3
-
-
i
i
1
1
4
0013
00I7
16207
0003
i
i
1
1
5
0011
0019
i
I
1
6
-
-
i
i
1
1
7
-
-
i
i
1
1
A
-
-?
0031
i
1
g
0042
0041
19473
i
I
1
10
i
i
1
1
ii
1&.686
i
1
1
12
0019
0028
0065
i
i
1
1
13
0002
0001
J
i
II
i
斗]ji
nnin
nme
in1匸斗
n斗qp
确定模型AR(3)MA(3)
GARCH模型建模
GARCH模型建模
DependentVariable:D(JPY}
MethodL已astSquares
Date-12/0B/17Time-12^5
Sample(阴usted『51427
Includedobservgtions:1423afteradjustments
Convergenceachievedafter3iterations
Variable
Coefficient
-Statistic
Prob
AR(1)
-2&
AR(2)
&
AR(3)
-00S7992
0.&2S425-3329931
0000s
R-5quared
Meandependentvar
AdjustedR-squared
[).010319
Akaikeinfoentenon
27&4471
Sumsquar-edresid
Schwarzcriterion
Loalikelihood
-19&
Hannan-Quinnenter
&3614
Durbin-Watsonstat
Inv/ertedARRoots
.26+351
,26-.35i-.47
AR⑴前面的系数显著为0,于是去掉ar⑴
DependentVariable:D(JPY)
Method:LeastSquares
Date:12/OB/17Time:12:4B
Sample(adjusted):51^127
neludedobservatians:1423afteradjustments
Donvergen匚亡achievedafter3iterations
Variable
Coefficient
-SLatistic
Prob.
AR(2)
0054145
0.&4M
AR(3)
-D085900
-
00012
R-squared
0009935
M&andependentvar
MList&dR-squared
:var
Akaikeinfocritenon
Sumsquaredresid
Schwarzcriterion
2762256
_oglikeliha口亡
-10&
Hannan-Quinncriter
Durbin-Watsonstat
nvertedARRoots
一24+.阳
,2^.35i-.4B
建立模型,参数显著性检验通过,残差是否是白噪声|~=]Equation:UWTITLEDWtarkfile:ARCHCASE12::Case1Z\-
[procfobject[Priit
Name^Freeze
Estimate
Forecast
StabReside
ofRtiiduali
view
Representations;
EstirriaticnOutput
t^
ARL1AStrucLure...
ARMAterms
Coeffi口mritDiagn口呂►
Rfi^idualDiagnostics►
StabilityDhagnosti匚s*►
Label
GradientsandDern/atives
CovarianceMatriK
ACPAGQ-StatProb
nnnA-
匚crrelogram-Q-staitistis...
CorrelDgramSquaredResiduals...
Histogram-IMormalhyTest
GARCH模型建模
GARCH模型建模
SerialCorrelationLMTest...
GARCH模型建模
GARCH模型建模
<i_*L4IIl|tfILim■IUN
□-statisticprobabilitiesadjustedfor2ARMAterms
Autuco(relation
PartialCorrelation
AC
PAC
Q-Stat
Prab
I
I
1
1
2OD01
-D001
27466
1
1
1
1
3-0002
-?
2751H
0097
1
1
1
1
40014
-1
&
1
1
1
1
1
1
1
6-
-D01^1
1
1
1
1
1-0007
-0005
33243
1
1
1
1
8-
■
O.€51
1
1
1
1
1
-
1
1
1
1
11-0D13
-
72700
□6Q9
1
1
1
120024
81360
D61B
1
1
1
1
13-
-
81429
1
1
1
1
14-0007-&
1
1
1
1
15ODIB
35701
D3Q6
(
1
r
1
16-0047
-DO49
11719
0629
1
1
1
12320
1
1
1
isn
nmo
<9讯
n了m
结果表明是白噪声序列,继续检验残差平方是否是自回归模型
ibjfct:SaveFrefi
lzjcIDet3ih+fJHhmv8Fmtizh1久口比llj~elf!tmli:3Enrl5aniDle!
Estimate
Forecast
Stats
Re&ids
1427-1427
1427-H27
ofResiduals
叵IEquation:UNTULEDV^rkfile:ARCHCASE12::Case12\[vlc»jPr-dc|objeciPrint|Name^reezs
Reprewntetions
EstimationOutput
CoeHideritDiagnaslics
ResidualDiagncities
►
StabililyDiagnoslics
卜
L沁I
ARKiAStructure...
GradienlsandDerivatives
CovansnceMgtris
)
ARMAterms
ACFACQ-StatProb
-
乃ACC打ACT匚4Ji卅
匸口rrelagrain-Q-statstics...
CorrelogrannSquaredResidual^,Histogram-hlurmalilyTestSerialCorrelationIMTesL.
MeteruskedasticityTests...
Date:12/08/17Time:12:48
Sample:11427
Includedobservations-1423
Autocorrelation
PartialCorrelatiDn
ACPACQ-StatProb
i
ZJ
i
ZJ
i
□
i
Zl
1
J
1
1
1
1
1
1
1
Zl
1
1
1
1
1
]
1
1
1
1
1
1
1
[
1
-
1
1
]
&7289.&
1
1
1
1
-
1
)
1
1
-
-
有自相关性,做ARCH效应检验
GARCH模型建模
GARCH模型建模
GARCH模型建模
GARCH模型建模
.-BinaryChoke(LaghsProbitiExtremeValue)
,=|Equation:UNTITLEDWorkfile:ARCHCA5E12::Case12\-nx
View|proc[Object][printjName
F『eezej
Estimate
Forecast
Stats
Re5id5
Representat0ns
ResidualsSquared
EstinnatitinOutput
A
ActuaLFi:tectReiidual
►
ARMAStructure...
GradientsandDerivsatives
k
in
ACPAC
Q-StatFrob
CovarianceMatrix
1
CoefficientDiagnostius
►
2
n>dncncrc
de-dTdinn
ResidualDiagnostics
Correlograrr-Q-&iati&tics...
StabilityDiagnostics
►
CarrelogramSquaredResiduals...
Label
Histogram-
NomnalityTest
-T」
"T1-
SerialCorrelaiion
LMTest...
1」
11
H^teroskedasticilyTests...
11
'1
1dd
nnann
□on0nnnn1
对话框选择ARCH效应
HetEfoskedasticityTest:ARCH
F-statistic
Prob
F(1p1420)
D0000
Obs^R-squaned
Prob.
Chi-Squane(1l)
TestEquation:
DependentVariableFTSID吃
Method:LeastSquares
Date:1ZW17Time:
Sample(adjusted)S1427
Includedobservations:1422afteradjustments
Vanable
Coetticient
t-Statistic
Prob.
C
&4
RESID'2(-1)
Ft-5quared
^4263
Meandependentvar
0917776
AdjustedR-squared
00^3604
「
2675652
S=ofregression
2589163
Akaikeinfocriterion
Sumsquaredresid
Schwarzcriterion
4749390
Loglikelihood
-
Hannan-Quinncriter.
4744716
F-statistic
□urbin-Watsonstat
2060666
口rCf匚g4p卜
会发现有ARCH效应,拒绝原假设(H0:al二a2二…二aq=0,即无条件异方差效应)
GARCH模型建模
GARCH模型建模
.-BinaryChoke(LaghsProbitiExtremeValue)
多次尝试,对残差平方定阶,选择7阶(PACF7阶截尾)
Date:12/08/17Time:12:48
Sample:11427
Includedobservations-1423
AutocorrelationPartialCorrelatiDnACPACQ-StatProb
ZJ
I
ZJ
□
I
Zl
:
1
1
1
1
1
Zl
1
1
1
]
1
1
1
]
[
1
]
1
]
)
1
1
)
1
1
01020028
-
-
-
.&
参数太多,考虑GARCH(1,1)模型
Sample..,
吧E12
Show...
-nX
1
Graph...
■口X
eFresz
EmptyGroup(Edit
CipnrSamDl?
□x
SeriesStatistics
GroupStaiisties
EstimateEquation...
EEtimata.
ObsfrR-squaned
|Estimate
Forecast
stats
Resids
^RCHCA5E1Z:;Case12\
(7p140fl)
JJ.,~rv~pT
-Square(7)
GARCH模型建模
GARCH模型建模
.-BinaryChoke(LaghsProbitiExtremeValue)
GARCH模型建模
GARCH模型建模
.-BinaryChoke(LaghsProbitiExtremeValue)
把均值模型和方差模型联立一起考虑
quatinnEiLtimatonx
勺pe询soonOptions
Equahcri叩已口亦曲叱品叮
DependentvariablefollowedhylistofregressorsincfludingiARMA
andPDLterms,ORanexplicitequationlikeV-c(l)+-c<2)TX.
dljoy)ar(2)ar(3)
semrtgsinLp
Method:
L£-Le酩t沱百(NlL£airtdARMA)
V
Sample;
LS-LeastSquares(NLSandARMAj
TSLS-Twa-StageLeastSquares:FSiNLSandARMA)
GMM-GeneralizedMethodofMoments
LIML-LimitedInforniatioriMaximumLikeliihoodlandK-Clas-&匚OINTTRK-CointegratiiigRegression
AR€H「山irtQTjg「£!53阳叱■€f»rdiliQneil14氏曰6皑!臼&5:如匚it
GARCH模型建模
GARCH模型建模
.-BinaryChoke(LaghsProbitiExtremeValue)