[R] empirical A matrix for dynamic system and markov analyses
Vassily Shvets
shv736 at yahoo.com
Tue Jul 12 10:39:34 CEST 2005
I'm working with a dynamic system that I've started to
analyse using msm(I've emailed to chris, orignator of
the program, separately, but maybe he's on holiday).
I'd like to know if anyone has considered the relation
between the markov analysis done with msm on the one
hand, and dynamic system analysis done with MATLAB?
That is, I'm working with a set of empirical data that
are more or less comparable to the kind of 'hospital'
analyses done in msm manual(my data has 4 states, and
'patients' move from one state to another over a time
period, n is varying from 120 to much larger).
Using the msm program gives me a nice Q and P matrices
for the system. However, I'm interested in making a
non homogenous system out of it(introducing
covariates, but covariates which also vary with time,
so that the system is approximately dy = Ax + Bz ).
One reason I'd like to go further than msm is taking
me right now is that the 'fit' statistic that msm
provides is I think, sort of narrowly focused on the
markov property. I guess I'm trying to ask if anyone
has analyzed such empirical data using the state space
system in MATLAB? The question also seems to involve
a/the ML estimation done by msm versus b/ the least
squares estimation of parameter that evidently is
predominant in the 'estimation toolbox' that runs with
MATLAB.
thanks,
s
More information about the R-help
mailing list