This was an example using package KFAS, which needs to be loaded before
running the code ('kf' is the function that computes Kalman filter in
package KFAS)
Giovanni
On Fri, 2011-08-26 at 07:11 -0700, quantguy wrote:
> Thanks! However, I added the line of code and receive an error during the
> opt
Thanks! However, I added the line of code and receive an error during the
optim() procedure:
rror in fn(par, ...) : could not find function "kf"
The update code is here:
tmp <- ts(read.table("http://shazam.econ.ubc.ca/intro/P.txt";,
header=T), start=c(1978,1), frequency=12) * 100
y <- tmp[,1:4
I just saw this old post, but it seems that nobody replied, so let me try.
If you can assume that also U[t] evelves as a random walk, I would build a
DLM by taking the state vector to be
x[t] = (U[t], UN[t], pi[t])'
By plugging in the equation for pi[t] the random walk expressions for U[t]
and
Oops..
You need to add the following line, right after the "m <- NCOL(y)"
statement:
k <- m * (m+1) / 2
'k' is the number of independent parameters in an m-by-m covariance
matrix, and two such matrices are estimated (Ht and Qt, both time
invariant). Hence the unknown parameter theta has length
I get the identical error even when applying the sample code from the dlm
vignette on state space models: http://www.jstatsoft.org/v41/i04/paper
The sample code is here:
tmp <- ts(read.table("http://shazam.econ.ubc.ca/intro/P.txt";, header=T),
start=c(1978,1), frequency=12) * 100
y <- tmp[,1:4] -
Thank you. The new article linked below is excellent. BTW, I have
had trouble getting the sample code to run.In particular, there is
a variable k that does not appear to be set when it is used at line
227.
Here is the line where it is called:
fit <- optim(par = rep(0, 2 * k), fn = logLik,
On Tue, 2011-06-07 at 17:24 +0100, Michael Ash wrote:
> This question pertains to setting up a model in the package "dlm"
> (dynamic linear models,
> http://cran.r-project.org/web/packages/dlm/index.html
The author of the dlm package has just published a paper on state space
models in R including
This question pertains to setting up a model in the package "dlm"
(dynamic linear models,
http://cran.r-project.org/web/packages/dlm/index.html
I have read both the vignette and "An R Package for Dynamic Linear
Models" (http://www.jstatsoft.org/v36/i12/paper), both of which are
very helpful. Ther
8 matches
Mail list logo
