Dear Ramani,
This is a state space formulation
Y_t= c_t + B_tX_t + e1
C_t=C_(t-1) +e2 e dis [0, V_t]
B_t = B_(t-1) + e3 ?
There are a few references and turorials under the title of Kalman
filtering, or more generally from a Bayesian perspective Bayesian
forecasting. You should find a few tutorials on the Internet if you look
under these keywords. You will also need to look for some software.
It might have been useful if you had said what you were trying to forecast
and why this model seemed appropriate.
Andrew
-----Original Message-----
From: A UK-based worldwide e-mail broadcast system mailing list
[mailto:[log in to unmask]]On Behalf Of RAMANI NATHAN
Sent: 09 September 2003 08:30
To: [log in to unmask]
Subject: advice
Dear Friends,
In my memory i did not send it earlier. This first time.
Mymodel yt = Dct + BXt + error
ct = ct-1 + error 2. Error and error 2 are
independent. This my multivariate time series non stationary model .
ct is the common trend and it is random walk. Xt is explanatory variable
and B is regression coefficient. yt is response variables.
I want to do the forecasting using this model. Can you advice me how
to do this? or relavent materials or web address. I am one research
student from developing country and in my country not enough materials.
Thank you.
Yours sincerely,
Ramani
---------------------------------
Post your free ad now! Yahoo! Canada Personals
|