Hi,

I'm trying to estimate parameters in the model:

y_t = y_{t-1}*u_{t}*lambda+N(0,sigma^2)

using Kalman filter. Where u{t} is t-th coordinate of the vector gathered from my data.
I use next space state model:

        eps_{t+1} = u{t} * lamba*eps_t + N(0,sigma^2)
        y_t = 1 * eps_t + N (0, sigma2^2)

I wrote the script to estimate parameters using mle, but it alway gives errors because mle can't find my vector U:

       
function void modify_F (matrix *F, scalar lambda, series *u, scalar t)
F[1,1]=lambda*u[t+1]
end function

kalman
obsy y
obsymat H
obsvar R
statemat F; modify_F(&F, lambda, &u, $kalman_t)
statevar Q
end kalman


mle logl = ERR ? NA : $kalman_llt

    Q[1,1]=sigma^2
    R[1,1]=sigma1^2
    ERR=kfilter()
    params sigma, lambda,  sigma1
end mle  --verbose --lbfgs

Is there any way how can I solve this problem?


Thanks,
Petrov Ivan

 
Privileged/Confidential Information may be contained in this message. If you
are not the addressee indicated in this message (or responsible for delivery
of the message to such person), you may not copy or deliver this message to 
anyone. In such case, you should destroy this message and kindly notify the
sender by reply email. Please advise immediately if you or your employer
does not consent to email for messages of this kind. Opinions, conclusions
and other information in this message that do not relate to the official
business of the GroupM companies shall be understood as neither given nor
endorsed by it.   GroupM companies are a member of WPP plc. For more
information on our business ethical standards and Corporate Responsibility 
policies please refer to our website at
http://www.wpp.com/WPP/About/