It is hard to say without the data. I suggest you to try the problem without the MLE part, such that you could check the state-space representation. For example, obtaining filtered data. Passing that test, it means that optimization is the problem. In that case, a good starting parameter-vector can be obtained from the associated ARIMA (or SARIMA) model.

2012/5/29 Paolo Chirico <paolo.chirico@unito.it>

I have run the following script for local linear model with quarterly seasonality:

scalar r = 10
scalar s1=1
scalar s2=1
scalar s3=1

matrix z = {1; 0; 1; 0; 0}

matrix Q = {s1^2,    0,    0,    0,    0;    0,    s2^2,    0,    0,    0;    0,    0,    s3^2,    0,    0;    0,    0,    0,    0,    0;    0,    0,    0,    0,    0}

matrix T = {1,    1,    0,    0,    0;    0,    1,    0,    0,    0;    0,    0,    -1,    -1,    -1;    0,    0,    1,    0,    0;    0,    0,    0,    1,    0}

kalman
    obsy y
    obsymat z
    statemat T
    obsvar r
    statevar Q
end kalman --diffuse

mle ll = ERR ? NA : $kalman_llt
    ERR = kfilter()
    params r s1 s2 s3
end mle

The model is a local linear model with quarterly seasonal.

But, I get the following error: failed to invert OPG matrix GG'

Where is the error? In the initial values of r, s1,..., in the script, or in kfilter()?

Best regards,
Paolo
-- Paolo Chirico Statistica Economica Dipartimento di Economia "S. Cognetti de Martiis" Università di Torino
_______________________________________________
Gretl-users mailing list
Gretl-users@lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users