The RATS Software Forum

Hi, i'm using GARCH(1,1) to make dynamic forecast, but I meet some problem when using Rats dafault coefficient

GARCH(P=1,Q=1) 2 501 ROU

GARCH Model - Estimation by BFGS
Convergence in 36 Iterations. Final criterion was 0.0000066 <= 0.0000100
Dependent Variable ROU
Usable Observations 500
Log Likelihood 822.6682

Variable Coeff Std Error T-Stat Signif
************************************************************************************
1. Mean 0.033548102 0.001517187 22.11204 0.00000000
2. C 0.000319291 0.000054336 5.87623 0.00000000
3. A 1.886544108 0.162348735 11.62032 0.00000000
4. B -0.002807887 0.003174761 -0.88444 0.37645864



AS YOU SEE, what is the default symbol of B and C?

likykiki wrote:Hi, i'm using GARCH(1,1) to make dynamic forecast, but I meet some problem when using Rats dafault coefficient

GARCH(P=1,Q=1) 2 501 ROU

GARCH Model - Estimation by BFGS
Convergence in 36 Iterations. Final criterion was 0.0000066 <= 0.0000100
Dependent Variable ROU
Usable Observations 500
Log Likelihood 822.6682

Variable Coeff Std Error T-Stat Signif
************************************************************************************
1. Mean 0.033548102 0.001517187 22.11204 0.00000000
2. C 0.000319291 0.000054336 5.87623 0.00000000
3. A 1.886544108 0.162348735 11.62032 0.00000000
4. B -0.002807887 0.003174761 -0.88444 0.37645864



AS YOU SEE, what is the default symbol of B and C?

It looks like you fit a GARCH model to a series where a GARCH doesn't really make sense.

TomDoan wrote:

likykiki wrote:Hi, i'm using GARCH(1,1) to make dynamic forecast, but I meet some problem when using Rats dafault coefficient

GARCH(P=1,Q=1) 2 501 ROU

GARCH Model - Estimation by BFGS
Convergence in 36 Iterations. Final criterion was 0.0000066 <= 0.0000100
Dependent Variable ROU
Usable Observations 500
Log Likelihood 822.6682

Variable Coeff Std Error T-Stat Signif
************************************************************************************
1. Mean 0.033548102 0.001517187 22.11204 0.00000000
2. C 0.000319291 0.000054336 5.87623 0.00000000
3. A 1.886544108 0.162348735 11.62032 0.00000000
4. B -0.002807887 0.003174761 -0.88444 0.37645864



AS YOU SEE, what is the default symbol of B and C?

It looks like you fit a GARCH model to a series where a GARCH doesn't really make sense.

Yeah, actually, my codes are:
open data o:\correlation.xlsx
data(format=xlsx,org=col) 2 3913 corr

set rou = corr**2.0

comp k=0
do i=501,3913,1

comp k=k+1

GARCH(P=1,Q=1,noprint) i-499 i ROU

comp b=%Beta(2)
comp c=%Beta(1)

My questions are : does beta (2) stands for B? and Beta(1) stands for C?

No. %beta(1) is the mean, %beta(2) is C, etc. Just as they are listed in the output.