Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Library: times
See also: nelmin

Quantlet: garchest
Description: estimates a GARCH process with mean zero by QMLE

Reference(s):

Usage: y = garchest(xx, q, p {, iniest, niniiter})
Input:
xx n x 1 vector, data
q scalar >= 1, number of lags in the ARCH part
p scalar >= 1, number of lags in the GARCH part
iniest optional (p+q+1) x m matrix, contains initial guesses of estimates; if not supplied, a random starting point is generated
niniiter optional scalar, number of randomly generated starting points (10 by default)
Output:
y list, consisting of components est, stderr, lik, s2
y.est (p+q+1) x 1 vector, parameter estimates
y.stderr (p+q+1) x 1 vector, standard errors
y.lik scalar, likelihood value
y.s2 n x 1 vector, volatility estimates

Example:
library("times")             ; loads the quantlets from times library
randomize(0)
dax=read("dax")             ; monthly DAX 1979:1-2000:10
daxreturn=tdiff(log(dax))   ; generates the monthly return
z=garchest(daxreturn,1,1)
z.est                        ; Parameter Estimates
z.stderr                     ; Standard Errors
z.lik                        ; Likelihood
dum = timeplot(z{4}, 261)   ; Plot of Volatility Estimate

Result:
Plots the estimated volatilities and produces the following output:

Contents of est
[1,]  0.00035762
[2,]  0.1267
[3,]  0.77079

Contents of stderr
[1,]  0.00026763
[2,]  0.065286
[3,]  0.11101

Contents of lik
[1,]   384.12
Example:
library("times")             ; loads the quantlets from times library
randomize(0)                 ; Seeding Initial Values
x=normal(100)                ; Generates normal distributed values
z=garchest(x,1,1)
z.est                        ; Parameter Estimates

Result:
Contents of est
[1,]  0.98828
[2,]  0.03317
[3,]  4.8934e-14



Author: P. Cizek, 20030522 license MD*Tech
(C) MD*TECH Method and Data Technologies, 05.02.2006