grad.dcc2 {ccgarch}R Documentation

Numerical gradient of the DCC part of the log-likelihood function

Description

This function computes numerical gradient of the second stage log-likelihood function of the DCC-GARCH model w.r.t. its parameters.

Usage

   grad.dcc2(param, dvar, d=1e-5)

Arguments

param a vector of DCC parameters (2 times 1)
dvar a matrix of standardised residuals (T times N)
d a step size for computing numerical gradient

Value

a vector of partial derivatives (2 times 1)

Note

The function is used only from dcc.estimation2 when the switch variable "gradient!=0". dlc$dlc in dlc contains the analytical gradient of the second stage log-likelihood function.

See Also

dlc dcc.estimation


[Package ccgarch version 0.1.1 Index]