autoarfima(data$VNI,ar.max = 2, ma.max = 2, criterion = "AIC", method = "full")
## $fit
##
## *----------------------------------*
## * ARFIMA Model Fit *
## *----------------------------------*
## Mean Model : ARFIMA(2,0,2)
## Distribution : norm
##
## Optimal Parameters
## ------------------------------------
## Estimate Std. Error t value Pr(>|t|)
## ar1 -0.599483 0.006708 -89.365 0
## ar2 -0.980294 0.005984 -163.825 0
## ma1 0.636549 0.002517 252.853 0
## ma2 1.003889 0.000055 18408.068 0
## sigma 0.013186 0.000477 27.648 0
##
## Robust Standard Errors:
## Estimate Std. Error t value Pr(>|t|)
## ar1 -0.599483 0.006997 -85.678 0
## ar2 -0.980294 0.005694 -172.172 0
## ma1 0.636549 0.002859 222.609 0
## ma2 1.003889 0.000056 18014.600 0
## sigma 0.013186 0.001025 12.860 0
##
## LogLikelihood : 1111.481
##
## Information Criteria
## ------------------------------------
##
## Akaike -5.7931
## Bayes -5.7415
## Shibata -5.7934
## Hannan-Quinn -5.7726
##
## Weighted Ljung-Box Test on Standardized Residuals
## ------------------------------------
## statistic p-value
## Lag[1] 0.8023 0.3704
## Lag[2*(p+q)+(p+q)-1][11] 4.1642 0.9997
## Lag[4*(p+q)+(p+q)-1][19] 6.1435 0.9647
##
## H0 : No serial correlation
##
## Weighted Ljung-Box Test on Standardized Squared Residuals
## ------------------------------------
## statistic p-value
## Lag[1] 9.46 2.100e-03
## Lag[2*(p+q)+(p+q)-1][2] 13.19 2.748e-04
## Lag[4*(p+q)+(p+q)-1][5] 29.04 5.911e-08
##
##
## ARCH LM Tests
## ------------------------------------
## Statistic DoF P-Value
## ARCH Lag[2] 14.45 2 7.290e-04
## ARCH Lag[5] 28.19 5 3.335e-05
## ARCH Lag[10] 32.26 10 3.618e-04
##
## Nyblom stability test
## ------------------------------------
## Joint Statistic: 1.5423
## Individual Statistics:
## ar1 0.32361
## ar2 0.07403
## ma1 0.04512
## ma2 0.16353
## sigma 1.08287
##
## Asymptotic Critical Values (10% 5% 1%)
## Joint Statistic: 1.28 1.47 1.88
## Individual Statistic: 0.35 0.47 0.75
##
##
## Elapsed time : 0.1024878
##
##
## $rank.matrix
## ar1 ar2 ma1 ma2 im arf AIC converged
## 1 1 1 1 1 0 0 -5.793095 1
## 2 1 1 1 1 1 0 -5.788216 1
## 3 1 1 0 0 0 0 -5.780487 1
## 4 0 1 1 0 0 0 -5.780412 1
## 5 0 1 0 0 0 0 -5.780048 1
## 6 1 0 0 1 0 0 -5.779575 1
## 7 0 0 1 1 0 0 -5.779357 1
## 8 0 0 0 1 0 0 -5.778982 1
## 9 0 0 1 0 0 0 -5.777081 1
## 10 0 1 1 1 0 0 -5.777070 1
## 11 0 1 0 1 0 0 -5.776783 1
## 12 1 1 0 1 0 0 -5.776521 1
## 13 1 0 0 0 0 0 -5.775890 1
## 14 1 1 0 0 1 0 -5.775731 1
## 15 0 1 1 0 1 0 -5.775649 1
## 16 0 1 0 0 1 0 -5.775391 1
## 17 1 1 1 0 0 0 -5.775258 1
## 18 1 0 1 0 0 0 -5.775185 1
## 19 1 0 0 1 1 0 -5.774819 1
## 20 0 0 1 1 1 0 -5.774591 1
## 21 1 0 1 1 0 0 -5.774371 1
## 22 0 0 0 1 1 0 -5.774320 1
## 23 0 1 1 1 1 0 -5.772310 1
## 24 0 0 1 0 1 0 -5.772209 1
## 25 0 1 0 1 1 0 -5.772099 1
## 26 1 1 0 1 1 0 -5.771772 1
## 27 0 0 0 0 1 0 -5.771412 1
## 28 1 0 0 0 1 0 -5.771028 1
## 29 1 1 1 0 1 0 -5.770500 1
## 30 1 0 1 0 1 0 -5.770334 1
## 31 1 0 1 1 1 0 -5.769621 1
autoarfima(data$SSEC,ar.max = 2, ma.max = 2, criterion = "AIC", method = "full")
## $fit
##
## *----------------------------------*
## * ARFIMA Model Fit *
## *----------------------------------*
## Mean Model : ARFIMA(2,0,2)
## Distribution : norm
##
## Optimal Parameters
## ------------------------------------
## Estimate Std. Error t value Pr(>|t|)
## mu -0.000288 0.000003 -99.854 0
## ar1 0.000000 NA NA NA
## ar2 0.967506 0.000993 974.154 0
## ma1 -0.062433 0.000877 -71.208 0
## ma2 -0.978199 0.000026 -37039.179 0
## sigma 0.009631 0.000307 31.377 0
##
## Robust Standard Errors:
## Estimate Std. Error t value Pr(>|t|)
## mu -0.000288 0.000006 -50.771 0
## ar1 0.000000 NA NA NA
## ar2 0.967506 0.002519 384.109 0
## ma1 -0.062433 0.001238 -50.435 0
## ma2 -0.978199 0.000058 -16943.645 0
## sigma 0.009631 0.000481 20.024 0
##
## LogLikelihood : 1224.791
##
## Information Criteria
## ------------------------------------
##
## Akaike -6.3863
## Bayes -6.3347
## Shibata -6.3867
## Hannan-Quinn -6.3659
##
## Weighted Ljung-Box Test on Standardized Residuals
## ------------------------------------
## statistic p-value
## Lag[1] 0.6246 0.4294
## Lag[2*(p+q)+(p+q)-1][11] 5.7122 0.6742
## Lag[4*(p+q)+(p+q)-1][19] 8.7438 0.6809
##
## H0 : No serial correlation
##
## Weighted Ljung-Box Test on Standardized Squared Residuals
## ------------------------------------
## statistic p-value
## Lag[1] 15.64 7.666e-05
## Lag[2*(p+q)+(p+q)-1][2] 32.32 2.614e-09
## Lag[4*(p+q)+(p+q)-1][5] 45.06 1.655e-12
##
##
## ARCH LM Tests
## ------------------------------------
## Statistic DoF P-Value
## ARCH Lag[2] 40.89 2 1.322e-09
## ARCH Lag[5] 43.42 5 3.030e-08
## ARCH Lag[10] 46.37 10 1.230e-06
##
## Nyblom stability test
## ------------------------------------
## Joint Statistic: 7.1162
## Individual Statistics:
## mu 0.06058
## ar2 0.06543
## ma1 0.05910
## ma2 0.05796
## sigma 1.29284
##
## Asymptotic Critical Values (10% 5% 1%)
## Joint Statistic: 1.28 1.47 1.88
## Individual Statistic: 0.35 0.47 0.75
##
##
## Elapsed time : 0.09291697
##
##
## $rank.matrix
## ar1 ar2 ma1 ma2 im arf AIC converged
## 1 0 1 1 1 1 0 -6.386342 1
## 2 0 0 0 1 0 0 -6.365746 1
## 3 0 1 0 0 0 0 -6.365457 1
## 4 0 0 0 0 1 0 -6.363651 1
## 5 0 0 1 0 0 0 -6.363403 1
## 6 1 0 0 0 0 0 -6.363381 1
## 7 1 1 0 1 0 0 -6.361419 1
## 8 0 0 0 1 1 0 -6.361038 1
## 9 0 1 0 0 1 0 -6.360741 1
## 10 1 0 0 1 0 0 -6.360700 1
## 11 0 0 1 1 0 0 -6.360661 1
## 12 1 1 0 0 0 0 -6.360463 1
## 13 0 1 1 0 0 0 -6.360429 1
## 14 0 1 0 1 0 0 -6.360018 1
## 15 1 1 1 0 0 0 -6.359862 1
## 16 1 0 1 1 0 0 -6.359833 1
## 17 1 0 1 0 0 0 -6.358907 1
## 18 0 0 1 0 1 0 -6.358635 1
## 19 1 0 0 0 1 0 -6.358614 1
## 20 1 1 0 1 1 0 -6.357031 1
## 21 0 1 0 1 1 0 -6.356980 1
## 22 0 1 1 1 0 0 -6.356538 1
## 23 1 0 0 1 1 0 -6.355980 1
## 24 0 0 1 1 1 0 -6.355940 1
## 25 1 1 0 0 1 0 -6.355735 1
## 26 0 1 1 0 1 0 -6.355701 1
## 27 1 1 1 0 1 0 -6.355379 1
## 28 1 0 1 1 1 0 -6.355352 1
## 29 1 1 1 1 0 0 -6.354646 1
## 30 1 0 1 0 1 0 -6.354136 1
## 31 1 1 1 1 1 0 -6.350176 1
CÁC DẠNG MÔ HÌNH CHO CHUỖI VNI
GJR-GARCH(11)VNI
vni.garch11n.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "norm")
vni.garch11n.fit <- ugarchfit(spec = vni.garch11n.spec, data = data$VNI)
vni.garch11t.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "std")
vni.garch11t.fit <- ugarchfit(spec = vni.garch11t.spec, data = data$VNI)
vni.garch11st.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sstd")
vni.garch11st.fit <- ugarchfit(spec = vni.garch11st.spec, data = data$VNI)
vni.garch11g.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "ged")
vni.garch11g.fit <- ugarchfit(spec = vni.garch11g.spec, data = data$VNI)
vni.garch11sg.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sged")
vni.garch11sg.fit <- ugarchfit(spec = vni.garch11sg.spec, data = data$VNI)
GJR-GARCH(12)VNI
vni.garch12n.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "norm")
vni.garch12n.fit <- ugarchfit(spec = vni.garch12n.spec, data = data$VNI)
vni.garch12t.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "std")
vni.garch12t.fit <- ugarchfit(spec = vni.garch12t.spec, data = data$VNI)
vni.garch12st.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sstd")
vni.garch12st.fit <- ugarchfit(spec = vni.garch12st.spec, data = data$VNI)
vni.garch12g.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "ged")
vni.garch12g.fit <- ugarchfit(spec = vni.garch12g.spec, data = data$VNI)
vni.garch12sg.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sged")
vni.garch12sg.fit <- ugarchfit(spec = vni.garch12sg.spec, data = data$VNI)
GJR-GARCH(21)VNI
vni.garch21n.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "norm")
vni.garch21n.fit <- ugarchfit(spec = vni.garch21n.spec, data = data$VNI)
vni.garch21t.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "std")
vni.garch21t.fit <- ugarchfit(spec = vni.garch21t.spec, data = data$VNI)
vni.garch21st.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sstd")
vni.garch21st.fit <- ugarchfit(spec = vni.garch21st.spec, data = data$VNI)
vni.garch21g.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "ged")
vni.garch21g.fit <- ugarchfit(spec = vni.garch21g.spec, data = data$VNI)
vni.garch21sg.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sged")
vni.garch21sg.fit <- ugarchfit(spec = vni.garch21sg.spec, data = data$VNI)
GJR-GARCH(22)VNI
vni.garch22n.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "norm")
vni.garch22n.fit <- ugarchfit(spec = vni.garch22n.spec, data = data$VNI)
vni.garch22t.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "std")
vni.garch22t.fit <- ugarchfit(spec = vni.garch22t.spec, data = data$VNI)
vni.garch22st.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sstd")
vni.garch22st.fit <- ugarchfit(spec = vni.garch22st.spec, data = data$VNI)
vni.garch22g.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "ged")
vni.garch22g.fit <- ugarchfit(spec = vni.garch22g.spec, data = data$VNI)
vni.garch22sg.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sged")
vni.garch22sg.fit <- ugarchfit(spec = vni.garch22sg.spec, data = data$VNI)
CÁC DẠNG MÔ HÌNH CHO CHUỖI SSEC
GJR-GARCH(11)SSEC
ssec.garch11n.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "norm")
ssec.garch11n.fit <- ugarchfit(spec = ssec.garch11n.spec, data = data$SSEC)
ssec.garch11t.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "std")
ssec.garch11t.fit <- ugarchfit(spec = ssec.garch11t.spec, data = data$SSEC)
ssec.garch11st.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sstd")
ssec.garch11st.fit <- ugarchfit(spec = ssec.garch11st.spec, data = data$SSEC)
ssec.garch11g.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "ged")
ssec.garch11g.fit <- ugarchfit(spec = ssec.garch11g.spec, data = data$SSEC)
ssec.garch11sg.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sged")
ssec.garch11sg.fit <- ugarchfit(spec = ssec.garch11sg.spec, data = data$SSEC)
GJR-GARCH(12)SSEC
ssec.garch12n.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "norm")
ssec.garch12n.fit <- ugarchfit(spec = ssec.garch12n.spec, data = data$SSEC)
ssec.garch12t.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "std")
ssec.garch12t.fit <- ugarchfit(spec = ssec.garch12t.spec, data = data$SSEC)
ssec.garch12st.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 2)),mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sstd")
ssec.garch12st.fit <- ugarchfit(spec = ssec.garch12st.spec, data = data$SSEC)
ssec.garch12g.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "ged")
ssec.garch12g.fit <- ugarchfit(spec = ssec.garch12g.spec, data = data$SSEC)
ssec.garch12sg.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(1, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sged")
ssec.garch12sg.fit <- ugarchfit(spec = ssec.garch12sg.spec, data = data$SSEC)
GJR-GARCH(21)SSEC
ssec.garch21n.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "norm")
ssec.garch21n.fit <- ugarchfit(spec = ssec.garch21n.spec, data = data$SSEC)
ssec.garch21t.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "std")
ssec.garch21t.fit <- ugarchfit(spec = ssec.garch21t.spec, data = data$SSEC)
ssec.garch21st.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sstd")
ssec.garch21st.fit <- ugarchfit(spec = ssec.garch21st.spec, data = data$SSEC)
ssec.garch21g.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "ged")
ssec.garch21g.fit <- ugarchfit(spec = ssec.garch21g.spec, data = data$SSEC)
ssec.garch21sg.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 1)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sged")
ssec.garch21sg.fit <- ugarchfit(spec = ssec.garch21sg.spec, data = data$SSEC)
GJR-GARCH(22)SSEC
ssec.garch22n.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "norm")
ssec.garch22n.fit <- ugarchfit(spec = ssec.garch22n.spec, data = data$SSEC)
ssec.garch22t.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "std")
ssec.garch22t.fit <- ugarchfit(spec = ssec.garch22t.spec, data = data$SSEC)
ssec.garch22st.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sstd")
ssec.garch22st.fit <- ugarchfit(spec = ssec.garch22st.spec, data = data$SSEC)
ssec.garch22g.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "ged")
ssec.garch22g.fit <- ugarchfit(spec = ssec.garch22g.spec, data = data$SSEC)
ssec.garch22sg.spec <- ugarchspec(variance.model = list(model = "gjrGARCH", garchOrder = c(2, 2)), mean.model = list(armaOrder = c(2, 2), include.mean = TRUE), distribution.model = "sged")
ssec.garch22sg.fit <- ugarchfit(spec = ssec.garch22sg.spec, data = data$SSEC)
LỰA CHỌN MÔ HÌNH GJR-GARCH
VNI
vni.model.list <- list(garch11n = vni.garch11n.fit, garch11t = vni.garch11t.fit, garch11st = vni.garch11st.fit, garch11g = vni.garch11g.fit, garch11sg = vni.garch11sg.fit, garch12n = vni.garch12n.fit, garch12t = vni.garch12t.fit, garch12st = vni.garch12st.fit, garch12g = vni.garch12g.fit, garch12sg = vni.garch12sg.fit, garch21n = vni.garch21n.fit, garch21t = vni.garch21t.fit, garch21st = vni.garch21st.fit, garch21g = vni.garch21g.fit, garch21sg = vni.garch21sg.fit, garch22n = vni.garch22n.fit, garch22t = vni.garch22t.fit, garch22st = vni.garch22st.fit, garch22g =vni.garch22g.fit, garch22sg = vni.garch22sg.fit)
vni.info.mat <- sapply(vni.model.list, infocriteria)
rownames(vni.info.mat) <- rownames(infocriteria(vni.garch11n.fit))
vni.info.mat
## garch11n garch11t garch11st garch11g garch11sg garch12n
## Akaike -6.013623 -6.057971 -6.061197 -6.066517 -6.062624 -6.011460
## Bayes -5.920668 -5.954688 -5.947585 -5.963233 -5.949013 -5.908177
## Shibata -6.014699 -6.059296 -6.062794 -6.067841 -6.064222 -6.012785
## Hannan-Quinn -5.976745 -6.016996 -6.016124 -6.025542 -6.017552 -5.970485
## garch12t garch12st garch12g garch12sg garch21n garch21t
## Akaike -6.059637 -6.055961 -6.071504 -6.057422 -6.012507 -6.060517
## Bayes -5.946025 -5.932021 -5.957893 -5.933482 -5.898895 -5.936577
## Shibata -6.061234 -6.057856 -6.073102 -6.059316 -6.014104 -6.062412
## Hannan-Quinn -6.014565 -6.006791 -6.026432 -6.008252 -5.967434 -6.011347
## garch21st garch21g garch21sg garch22n garch22t garch22st
## Akaike -6.054132 -6.064112 -6.060230 -6.010826 -6.055281 -6.051581
## Bayes -5.919864 -5.940172 -5.925962 -5.886886 -5.921013 -5.906984
## Shibata -6.056349 -6.066006 -6.062446 -6.012721 -6.057498 -6.054143
## Hannan-Quinn -6.000865 -6.014942 -6.006962 -5.961656 -6.002014 -5.994216
## garch22g garch22sg
## Akaike -6.058876 -6.054994
## Bayes -5.924608 -5.910398
## Shibata -6.061092 -6.057556
## Hannan-Quinn -6.005609 -5.997629
vni.inds <- which(vni.info.mat == min(vni.info.mat), arr.ind=TRUE)
model.vni <- colnames(vni.info.mat)[vni.inds[,2]]
model.vni
## [1] "garch12g"
SSEC
ssec.model.list <- list(garch11n = ssec.garch11n.fit, garch11t = ssec.garch11t.fit, garch11st = ssec.garch11st.fit, garch11g = ssec.garch11g.fit, garch11sg = ssec.garch11sg.fit, garch12n = ssec.garch12n.fit, garch12t = ssec.garch12t.fit, garch12st = ssec.garch12st.fit, garch12g = ssec.garch12g.fit, garch12sg = ssec.garch12sg.fit, garch21n = ssec.garch21n.fit, garch21t = ssec.garch21t.fit, garch21st = ssec.garch21st.fit, garch21g = ssec.garch21g.fit, garch21sg = ssec.garch21sg.fit, garch22n = ssec.garch22n.fit, garch22t = ssec.garch22t.fit, garch22st = ssec.garch22st.fit, garch22g =ssec.garch22g.fit, garch22sg = ssec.garch22sg.fit)
ssec.info.mat <- sapply(ssec.model.list, infocriteria)
rownames(ssec.info.mat) <- rownames(infocriteria(ssec.garch11n.fit))
ssec.info.mat
## garch11n garch11t garch11st garch11g garch11sg garch12n
## Akaike -6.472143 -6.512189 -6.492636 -6.497560 -6.506762 -6.466909
## Bayes -6.379188 -6.408906 -6.379024 -6.394277 -6.393151 -6.363626
## Shibata -6.473219 -6.513514 -6.494233 -6.498884 -6.508360 -6.468233
## Hannan-Quinn -6.435265 -6.471214 -6.447563 -6.456585 -6.461690 -6.425934
## garch12t garch12st garch12g garch12sg garch21n garch21t
## Akaike -6.486715 -6.480882 -6.476106 -6.490608 -6.461674 -6.484157
## Bayes -6.373103 -6.356942 -6.362495 -6.366668 -6.348063 -6.360217
## Shibata -6.488312 -6.482777 -6.477704 -6.492502 -6.463271 -6.486052
## Hannan-Quinn -6.441642 -6.431712 -6.431034 -6.441438 -6.416602 -6.434987
## garch21st garch21g garch21sg garch22n garch22t garch22st
## Akaike -6.487089 -6.480626 -6.502148 -6.501030 -6.479709 -6.445446
## Bayes -6.352821 -6.356686 -6.367880 -6.377090 -6.345441 -6.300850
## Shibata -6.489305 -6.482521 -6.504364 -6.502924 -6.481925 -6.448008
## Hannan-Quinn -6.433822 -6.431456 -6.448881 -6.451860 -6.426442 -6.388081
## garch22g garch22sg
## Akaike -6.486389 -6.469164
## Bayes -6.352120 -6.324567
## Shibata -6.488605 -6.471726
## Hannan-Quinn -6.433121 -6.411799
ssec.inds <- which(ssec.info.mat == min(ssec.info.mat), arr.ind=TRUE)
model.ssec <- colnames(ssec.info.mat)[ssec.inds[,2]]
model.ssec
## [1] "garch11t"
THAM SỐ ƯỚC LƯỢNG MÔ HÌNH BIÊN PHÙ HỢP NHẤT
vni.garch12g.fit
##
## *---------------------------------*
## * GARCH Model Fit *
## *---------------------------------*
##
## Conditional Variance Dynamics
## -----------------------------------
## GARCH Model : gjrGARCH(1,2)
## Mean Model : ARFIMA(2,0,2)
## Distribution : ged
##
## Optimal Parameters
## ------------------------------------
## Estimate Std. Error t value Pr(>|t|)
## mu -0.001101 0.000366 -3.006446 0.002643
## ar1 -0.613299 0.008161 -75.146825 0.000000
## ar2 -0.981956 0.008741 -112.345010 0.000000
## ma1 0.638120 0.005069 125.881951 0.000000
## ma2 0.996587 0.002644 376.885130 0.000000
## omega 0.000004 0.000008 0.459440 0.645918
## alpha1 0.049165 0.018161 2.707131 0.006787
## beta1 0.863481 0.726343 1.188806 0.234516
## beta2 0.000001 0.616319 0.000001 0.999999
## gamma1 0.172706 0.073491 2.350049 0.018771
## shape 1.209651 0.116444 10.388224 0.000000
##
## Robust Standard Errors:
## Estimate Std. Error t value Pr(>|t|)
## mu -0.001101 0.000327 -3.36752 0.000758
## ar1 -0.613299 0.008392 -73.08333 0.000000
## ar2 -0.981956 0.007283 -134.82571 0.000000
## ma1 0.638120 0.008158 78.21679 0.000000
## ma2 0.996587 0.003860 258.15806 0.000000
## omega 0.000004 0.000022 0.16223 0.871122
## alpha1 0.049165 0.097383 0.50486 0.613655
## beta1 0.863481 2.454360 0.35181 0.724977
## beta2 0.000001 2.125623 0.00000 1.000000
## gamma1 0.172706 0.364297 0.47408 0.635442
## shape 1.209651 0.111046 10.89324 0.000000
##
## LogLikelihood : 1170.657
##
## Information Criteria
## ------------------------------------
##
## Akaike -6.0715
## Bayes -5.9579
## Shibata -6.0731
## Hannan-Quinn -6.0264
##
## Weighted Ljung-Box Test on Standardized Residuals
## ------------------------------------
## statistic p-value
## Lag[1] 0.04154 0.8385
## Lag[2*(p+q)+(p+q)-1][11] 3.91600 1.0000
## Lag[4*(p+q)+(p+q)-1][19] 5.93752 0.9732
## d.o.f=4
## H0 : No serial correlation
##
## Weighted Ljung-Box Test on Standardized Squared Residuals
## ------------------------------------
## statistic p-value
## Lag[1] 1.308 0.2527
## Lag[2*(p+q)+(p+q)-1][8] 3.434 0.6060
## Lag[4*(p+q)+(p+q)-1][14] 6.040 0.6335
## d.o.f=3
##
## Weighted ARCH LM Tests
## ------------------------------------
## Statistic Shape Scale P-Value
## ARCH Lag[4] 0.004419 0.500 2.000 0.9470
## ARCH Lag[6] 0.978788 1.461 1.711 0.7538
## ARCH Lag[8] 2.880249 2.368 1.583 0.5661
##
## Nyblom stability test
## ------------------------------------
## Joint Statistic: 4.7032
## Individual Statistics:
## mu 0.58628
## ar1 0.03408
## ar2 0.03953
## ma1 0.02026
## ma2 0.09719
## omega 0.69342
## alpha1 0.14404
## beta1 0.18409
## beta2 0.18419
## gamma1 0.26092
## shape 0.01994
##
## Asymptotic Critical Values (10% 5% 1%)
## Joint Statistic: 2.49 2.75 3.27
## Individual Statistic: 0.35 0.47 0.75
##
## Sign Bias Test
## ------------------------------------
## t-value prob sig
## Sign Bias 0.1981 0.8431
## Negative Sign Bias 0.4439 0.6573
## Positive Sign Bias 1.0941 0.2746
## Joint Effect 2.0079 0.5708
##
##
## Adjusted Pearson Goodness-of-Fit Test:
## ------------------------------------
## group statistic p-value(g-1)
## 1 20 16.95 0.5931
## 2 30 27.32 0.5545
## 3 40 52.35 0.0749
## 4 50 53.34 0.3110
##
##
## Elapsed time : 0.71386
ssec.garch11t.fit
##
## *---------------------------------*
## * GARCH Model Fit *
## *---------------------------------*
##
## Conditional Variance Dynamics
## -----------------------------------
## GARCH Model : gjrGARCH(1,1)
## Mean Model : ARFIMA(2,0,2)
## Distribution : std
##
## Optimal Parameters
## ------------------------------------
## Estimate Std. Error t value Pr(>|t|)
## mu 0.000003 0.000000 7.8696e+01 0.000000
## ar1 1.678839 0.000015 1.1487e+05 0.000000
## ar2 -0.689246 0.000022 -3.0647e+04 0.000000
## ma1 -1.649835 0.000107 -1.5459e+04 0.000000
## ma2 0.640238 0.000061 1.0452e+04 0.000000
## omega 0.000009 0.000000 4.2422e+01 0.000000
## alpha1 0.000000 0.000007 1.1640e-02 0.990713
## beta1 0.796221 0.023351 3.4098e+01 0.000000
## gamma1 0.221325 0.062582 3.5366e+00 0.000405
## shape 9.310836 1.164876 7.9930e+00 0.000000
##
## Robust Standard Errors:
## Estimate Std. Error t value Pr(>|t|)
## mu 0.000003 0.000002 1.0506e+00 0.293459
## ar1 1.678839 0.000286 5.8746e+03 0.000000
## ar2 -0.689246 0.000435 -1.5850e+03 0.000000
## ma1 -1.649835 0.007448 -2.2150e+02 0.000000
## ma2 0.640238 0.007494 8.5430e+01 0.000000
## omega 0.000009 0.000006 1.6519e+00 0.098557
## alpha1 0.000000 0.000162 4.7300e-04 0.999622
## beta1 0.796221 2.093264 3.8037e-01 0.703669
## gamma1 0.221325 5.446085 4.0639e-02 0.967583
## shape 9.310836 54.878383 1.6966e-01 0.865275
##
## LogLikelihood : 1253.828
##
## Information Criteria
## ------------------------------------
##
## Akaike -6.5122
## Bayes -6.4089
## Shibata -6.5135
## Hannan-Quinn -6.4712
##
## Weighted Ljung-Box Test on Standardized Residuals
## ------------------------------------
## statistic p-value
## Lag[1] 1.830 0.1762
## Lag[2*(p+q)+(p+q)-1][11] 5.507 0.7903
## Lag[4*(p+q)+(p+q)-1][19] 10.687 0.3591
## d.o.f=4
## H0 : No serial correlation
##
## Weighted Ljung-Box Test on Standardized Squared Residuals
## ------------------------------------
## statistic p-value
## Lag[1] 0.4691 0.4934
## Lag[2*(p+q)+(p+q)-1][5] 2.9462 0.4170
## Lag[4*(p+q)+(p+q)-1][9] 5.2369 0.3954
## d.o.f=2
##
## Weighted ARCH LM Tests
## ------------------------------------
## Statistic Shape Scale P-Value
## ARCH Lag[3] 2.630 0.500 2.000 0.1048
## ARCH Lag[5] 3.774 1.440 1.667 0.1955
## ARCH Lag[7] 5.328 2.315 1.543 0.1934
##
## Nyblom stability test
## ------------------------------------
## Joint Statistic: 27.2051
## Individual Statistics:
## mu 5.03682
## ar1 0.04385
## ar2 0.04982
## ma1 0.04425
## ma2 0.04954
## omega 3.41948
## alpha1 1.10684
## beta1 0.51082
## gamma1 1.08609
## shape 0.30814
##
## Asymptotic Critical Values (10% 5% 1%)
## Joint Statistic: 2.29 2.54 3.05
## Individual Statistic: 0.35 0.47 0.75
##
## Sign Bias Test
## ------------------------------------
## t-value prob sig
## Sign Bias 0.3052 0.7604
## Negative Sign Bias 0.5008 0.6168
## Positive Sign Bias 0.9271 0.3545
## Joint Effect 1.5159 0.6786
##
##
## Adjusted Pearson Goodness-of-Fit Test:
## ------------------------------------
## group statistic p-value(g-1)
## 1 20 14.96 0.7249
## 2 30 26.06 0.6221
## 3 40 30.57 0.8307
## 4 50 31.87 0.9724
##
##
## Elapsed time : 0.7852011
KIỂM ĐỊNH SỰ PHÙ HỢP CỦA MÔ HÌNH BIÊN
Trích xuất chuỗi phần dư u của chuỗi SSEC
ssec.res <- residuals(ssec.garch11t.fit)/sigma(ssec.garch11t.fit)
fitdist(distribution = "std", ssec.res, control = list())
## $pars
## mu sigma shape
## -0.04440066 0.96483477 14.81873604
##
## $convergence
## [1] 0
##
## $values
## [1] 534.8834 526.8599 526.8599
##
## $lagrange
## [1] 0
##
## $hessian
## [,1] [,2] [,3]
## [1,] 476.4714363 -6.9584771 0.152151116
## [2,] -6.9584771 692.8693079 0.527188118
## [3,] 0.1521511 0.5271881 0.007650928
##
## $ineqx0
## NULL
##
## $nfuneval
## [1] 114
##
## $outer.iter
## [1] 2
##
## $elapsed
## Time difference of 0.02116299 secs
##
## $vscale
## [1] 1 1 1 1
u <- pdist(distribution = "std", q = ssec.res, mu = -0.04440066 , sigma = 0.96483477 , shape= 14.81873604)
Trích xuất chuỗi phần dư v của chuỗi VNI
vni.res <- residuals(vni.garch12g.fit)/sigma(vni.garch12g.fit)
fitdist(distribution = "ged", vni.res, control = list())
## $pars
## mu sigma shape
## 0.0270057 1.0160627 1.2064201
##
## $convergence
## [1] 0
##
## $values
## [1] 547.7608 534.9186 534.9186
##
## $lagrange
## [1] 0
##
## $hessian
## [,1] [,2] [,3]
## [1,] 370.62020 10.64419 -26.01218
## [2,] 10.64419 462.89619 46.65337
## [3,] -26.01218 46.65337 76.03417
##
## $ineqx0
## NULL
##
## $nfuneval
## [1] 71
##
## $outer.iter
## [1] 2
##
## $elapsed
## Time difference of 0.02176595 secs
##
## $vscale
## [1] 1 1 1 1
v <- pdist("ged",vni.res, mu = 0.0270057, sigma = 1.0160627 , shape = 1.2064201)
Các kiểm định sự phù hợp mô hình biên
Kiểm định Anderson-Darling (A-D)
ad.test(u, "punif")
##
## Anderson-Darling test of goodness-of-fit
## Null hypothesis: uniform distribution
## Parameters assumed to be fixed
##
## data: u
## An = 0.31901, p-value = 0.9232
ad.test(v, "punif")
##
## Anderson-Darling test of goodness-of-fit
## Null hypothesis: uniform distribution
## Parameters assumed to be fixed
##
## data: v
## An = 0.35288, p-value = 0.8938
Kiểm định Cramer-von Mises (Cv-M)
cvm.test(u, "punif")
##
## Cramer-von Mises test of goodness-of-fit
## Null hypothesis: uniform distribution
## Parameters assumed to be fixed
##
## data: u
## omega2 = 0.045375, p-value = 0.904
cvm.test(v, "punif")
##
## Cramer-von Mises test of goodness-of-fit
## Null hypothesis: uniform distribution
## Parameters assumed to be fixed
##
## data: v
## omega2 = 0.03482, p-value = 0.958
Kiểm định Kolmogorov-Smornov (K-S).
ks.test(u, "punif")
##
## Asymptotic one-sample Kolmogorov-Smirnov test
##
## data: u
## D = 0.027447, p-value = 0.9358
## alternative hypothesis: two-sided
ks.test(v, "punif")
##
## Asymptotic one-sample Kolmogorov-Smirnov test
##
## data: v
## D = 0.028023, p-value = 0.9251
## alternative hypothesis: two-sided
ƯỚC LƯỢNG THAM SỐ COUPULA
BiCopSelect(u, v, familyset= NA, selectioncrit="AIC",indeptest = FALSE, level = 0.05)
## Bivariate copula: Gumbel (par = 1.13, tau = 0.11)
Stu <- BiCopEst(u, v, family = 4, method = "mle", se = T, max.df = 10)
summary(Stu)
## Family
## ------
## No: 4
## Name: Gumbel
##
## Parameter(s)
## ------------
## par: 1.13 (SE = 0.04)
##
## Dependence measures
## -------------------
## Kendall's tau: 0.11 (empirical = 0.13, p value < 0.01)
## Upper TD: 0.15
## Lower TD: 0
##
## Fit statistics
## --------------
## logLik: 7.49
## AIC: -12.97
## BIC: -9.03