Package "dlm" has a function for maximum likelihood estimation of parameters in general (linear Normal) state space model. The function, dlmMLE, computes the likelihood based on singular value decomposition and appears to be fairly robust.
No EM algorithm, though. Giovanni > Date: Sun, 11 Nov 2007 18:32:32 +0000 (GMT) > From: [EMAIL PROTECTED] > Sender: [EMAIL PROTECTED] > Priority: normal > Precedence: list > > Hi - I follow some references and now implement my own state-space model > estimation. I have a question. In case, my equations are like this: > > y(t) = Ax(t)+Bu(t)+eps(t) # observation eq > x(t) = Cx(t-1)+Du(t)+eta(t) # state eq > > Using EM, after backward recursion, you will use the smoothed state > estimation to update the A, B, C, and D which is chosen so as to maximize the > expectation equation. But for example, if my C is always of matrix zero > (model specification), during EM I still get the value of estimated C which > turns out to be non-zero. How can I resolve this conflict? Or I just ignore > my estimation result and keep it as zero? THank you. > > - adschai > > ______________________________________________ > R-help@r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. > > -- Giovanni Petris <[EMAIL PROTECTED]> Associate Professor Department of Mathematical Sciences University of Arkansas - Fayetteville, AR 72701 Ph: (479) 575-6324, 575-8630 (fax) http://definetti.uark.edu/~gpetris/ ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
