estim_ebee: Estimation of multivariate GARCH CCC equation by equation

In TaoussD/EbEEMGARCH: Estimating MGARCH(1,1) model equation by equation

Description

Estimation of a MGARCH(1,1) CCC-diagonal or semi-diagonal equation by equation.

Usage

estimCCC.EbEE(Omega, Alpha, Beta, eps, r = 10, model)

Arguments

Omega	Initialisation parameter for Omega (vector)
Alpha	Initialisation parameter for Alpha (vector), must be a matrix if you are using the "sdiagonal" model
Beta	Initialisation parameter for Beta (vector)
eps	m*n matrix of the serie to be estimated
r	Number of observations used for the initialisation
model	model="diagonal" if MGARCH(1,1) diagonal model="sdiagonal" if MGARCH(1,1) is semi-diagonal i.e. Alpha is plain and Beta diagonal

Value

Omega	Estimation of Omega
Alpha	Estimation of Alpha
Beta	Estimation of Beta
R	Estimation of R, the correlation matrix
Residuals	Estimation of the residuals

Author(s)

D. Taouss, C. Francq

References

C. Francq and J.M. Zakoian, Estimating multivariate GARCH and Stochastic Correlation models equation by equation, October 2014

See Also

EbEEMGARCH Homepage of the documentation

Examples

####
# Sampling a diagonal model
####

m <- 3 #Number of assets in the portfolio
Omega0 <- rep(0.01, m)
Alpha0 <- rep(0.05, m)
Beta0 <- rep(0.90, m)
R0 <- diag(rep(1, m))

Epsi <- GarchCCC.sim(2500, Omega0, Alpha0, Beta0,"diagonal", R0, "normal")


####
# Estimation of the parameters
####

Omegainit <- rep(0.1, m)
Alphainit <- rep(0.5, m)
Betainit <- rep(0.7, m)

estimCCC.EbEE(Omegainit,Alphainit,Betainit,Epsi,r=10,model="diagonal")

####
# If we allow a semi-diagonal model
####

omegainit <- rep(0.2, m)
alphainit <- matrix(rep(0.2 / m, m ^ 2), nrow = m)
betainit <- rep(0.70, m)

estimCCC.EbEE(omegainit,alphainit,betainit,Epsi,r=10,model="sdiagonal")

TaoussD/EbEEMGARCH documentation built on May 9, 2019, 4:18 p.m.