state space model with arch errors

[Yohan]

Hello altogether,

i want to program a state space model with arch errors.

For that purpose, i need - for the error e - a error variance formula of the following form:

Var(e)=E(e^2)=c(1)+c(2)*E(e^2(-1))

That implies, i need the residuals e^2(-1) of the day before in the estimation procedure.

HOW CAN I STORE AND USE THE RESIDUALS OF THE STATE SPACE MODEL DURING THE ML ESTIMATION???

So far, i have only a idea to select the residuals AFTER the estimation procedure, as follows:

uoc.append @state sv1=sv1(-1) + e2
uoc.append @state sv2=sv2(-1)
uoc.append @state sv3=sv3(-1) + e4
uoc.append @state sv4=sv4(-1)
uoc.append @state sv5=sv5(-1) + e6

uoc.append @signal cds=sv1 + sv2*aktie(-1) + sv3*implied_volatility(-1) +sv4*swap(-1) +sv5*cds(-1)

uoc.ml

series se1
series se2
series se3
series se4
series se5
series se6

uoc.makesignals(t=disturb) se1
uoc.makestates(t=disturb) se2 se3 se4 se5 se6

THANK YOU VERY MUCH FOR YOUR ANSWERS!!!

Best regards, Yohan