similar to: The constant part of the log-likelihood in StructTS

I am trying to make the following Kalman filter equations work and therefore produce their graphs. v_t=y_t - a_t a_t+1=a_t+K_t*v_t F_t=P_t+sigma.squared.epsilon P_t+1=P_t*(1-K_t)+sigma.squared.eta K_t=P_t/F_t Given: a_1=0,P_1=10^7,sigma.squared.epsilon=15099, sigma.squared.eta=1469.1 I have attached my code,which of course doesnt work.It produces NAs for the Fs,Ks and the a. Can somebody tell me

Dear all, I was playing around with factor contrasts, and found the argument nmax on function factor. When using nmax=1, R froze completely, and I had to close it from task manager. After some debugging, I found that the problem is actually in unique-function, where the internal unique function is called: .Internal(unique(x, incomparables, fromLast, nmax)) More generally, it looks like

Dear All, I am trying to fit a structural time series model using the StructTS function (package stats) with only 2 seasons (summer and winter). More than 2 seasons work fine but with 2 seasons I get this error: > fit <- StructTS(y.ts, type="BSM") Error in T[cbind(ind + 1L, ind)] <- 1 : subscript out of bounds I have looked at Prof. Ripley's 2002 RNews article but cannot

To whomever it may concern, I'm a young Industrial Engineer working on Senior Design at Georgia Tech and have found the StructTS method to be excellent for the training set for my forecasting project. There's only one problem: I don't actually understand what a Structural Time Series IS. I've looked up resources on it, and get that essentially you're dividing the Time

Dear all, The definition of the initial covariance matrix P in StructTS function seems to be defined in a somewhat non-standard way without any references. Usually that matrix is defined as a diagonal matrix in case of structural time series models, but StructTS defines this as a singular matrix filled with 1e+06 * var(x, na.rm = TRUE)/100 where x is the time series being modelled. I wonder if

I apologize in advance if I picked the wrong list to post this to. I have made an effort to find the answers to these questions on CRAN, but if they are there, I couldn't find them, and I was going to email the developer of StructTS directly but could not find who that is. I have 2 interrelated questions about StructTS 1. Where can I obtain the source code for StructTS if I wanted to

Hi all, I've experienced this problem using StructTS(x) when the *first* element of x is a NA (R:R1.7.0, os: w2ksp4). Please look at the following code: a=rep(1:7,10) library(ts) #this works StructTS(a) #this works x=a x[2]=NA StructTS(x) #this doesn't work x=a x[1]=NA StructTS(x) The last command returns this error "Error in optim(init[mask], getLike,

Dear all, I am applying the StructTS function in ts-package. For some time series the program terminates and the following error appears: Error in optim(init[mask], getLike, method = "L-BFGS-B", lower = rep(0, : L-BFGS-B needs finite values of fn Do someone know what do I have to adjust in the original time series to avoid this error? It works fine for some subsets of the

Hi, I used function StructTS some time ago to fit a structural model to a time series. Now with R 1.6.2-1 I repeated the analysis with the same series and I get the following error: Error in KalmanLike2(y, Z, -1) : invalid argument type I tried with other series and I get the same error; I checked the examples in the documentation and they work fine. I suspect I am missing something

Hello, Does anybody know if (and how) it is possible to obtain standard errors of estimated variances from StructTS? (R 2.10.0). Thank you in advance, Giovanni

Dear R-bugs I have been studying the StructTS function (in package 'stats') and functions supplied with it. I think I have found a few minor bugs in the documentation. I am referring to the version of StructTS supplied with the release R 2.0.0. Output from 'version' platform i386-pc-mingw32 arch i386 os mingw32 system i386, mingw32 status

In the examples for StructTS -- ($RHOME)/library/stats/man/StructTS.Rd -- could par(mfrow = c(4, 1)) plot(log10(UKgas)) plot(cbind(fitted(fit), resids=resid(fit)), main = "UK gas consumption") become plot(log10(UKgas)) par(mfrow = c(4, 1)) plot(cbind(fitted(fit), resids=resid(fit)), main = "UK gas consumption") ## Note that par was moved down This makes the plot of UKgas

Dear All I would like to use the StructTS function in the ts library to fit the ' BSM ' model. I have some, probably basics, questions about the model and about the function(s): 1) How can I check the statistical significance of the estimated parameters(variances)? 2) Is there some way to find what component "dominate" the series? 3) Is there a function to produce

(re-formulate, re-send, without html) for vector y = c(1,2,3,4,5), H = 0.66 manual calculations using the equations below give a = c(1,1.66,2.55,3.51,4.50). KalmanRun with these parameters gives res$states = (1,1,1,1,1)! for Kalman Filter Durbin/Koopman give at p67 eqs 4.13: v = y - Z a, F = Z P Z' + H, K = T P Z' / F + H, a[t+1] = T a + K v, P[t+1] = T P L'

1. The function "KalmanLike" seems to change its inputs AND PREVIOUSLY MADE copies of the inputs. Consider the following (using R 2.1.0 patched under Windows XP): > Fig2.1 <- StructTS(x=Nile, type="level") > unlist(Fig2.1$model0[2:3]) a P 1120 286379470 > tst2 <- tst <- Fig2.1$model0 > tst23 <- tst[2:3] > tst23u <-

Hi, I am trying to figure how to use dlm with constant terms (possibly time-dependent) added to both equations y_t = c_t + F_t\theta_t + v_t \theta_t = d_t + G_t\theta_{t-1} + w_t, in the way that S-PLUS Finmetrics does? Is there any straightforward way to transform the above to the default setup? Thanks, Tsvetan -------------------------------------------------------- NOTICE: If received in

Dear All I would like to use the StructTS function in the ts library to fit the ' BSM ' model. I have some, probably basics, questions about the model and about the function(s): 1) How can I check the statistical significance of the estimated parameters(variances)? 2) Is there some way to find what component "dominate" the series? 3) Is there a function to produce

This is only partly a question about R, as I am not quite sure about the underlying statistical theory either. I have fitted a non-linear mixed-effects model with nlme. In the fixed part of the model I have a factor with three levels as explanatory variable. I would like to use Tukey HSD or a similar test to test for differences between these three levels. I have two grouping factors:

JSS, which is at http://www.jstatsoft.org, publishes statistical software with manuals, as well as papers describing, comparing, and evaluating statistical software. Abstracts of all contributions accepted in JSS are published in printed form in JCGS. Although we publish software written in any language, we especially welcome software written in R. The general idea is that packages written in R

JSS, which is at http://www.jstatsoft.org, publishes statistical software with manuals, as well as papers describing, comparing, and evaluating statistical software. Abstracts of all contributions accepted in JSS are published in printed form in JCGS. Although we publish software written in any language, we especially welcome software written in R. The general idea is that packages written in R