[Pomp-commits] r51 - in pkg: R man tests
noreply at r-forge.r-project.org
noreply at r-forge.r-project.org
Mon Aug 25 23:23:02 CEST 2008
Author: kingaa
Date: 2008-08-25 23:23:02 +0200 (Mon, 25 Aug 2008)
New Revision: 51
Modified:
pkg/R/mif.R
pkg/man/mif.Rd
pkg/tests/ou2-mif.Rout.save
Log:
error trap in 'mif': 'rw.sd' must have names
Modified: pkg/R/mif.R
===================================================================
--- pkg/R/mif.R 2008-08-25 17:14:21 UTC (rev 50)
+++ pkg/R/mif.R 2008-08-25 21:23:02 UTC (rev 51)
@@ -77,6 +77,8 @@
)
rw.names <- names(rw.sd)
+ if (is.null(rw.names))
+ stop("mif error: ",sQuote("rw.sd")," must be a named vector",call.=FALSE)
if (any(!(rw.names%in%start.names)))
stop("mif error: all the names of ",sQuote("rw.sd")," must be names of ",sQuote("start"),call.=FALSE)
if (any(rw.sd[c(pars,ivps)]<=0)) {
Modified: pkg/man/mif.Rd
===================================================================
--- pkg/man/mif.Rd 2008-08-25 17:14:21 UTC (rev 50)
+++ pkg/man/mif.Rd 2008-08-25 21:23:02 UTC (rev 51)
@@ -44,7 +44,7 @@
If \code{particles} is not supplied by the user, the default behavior is to draw the particles from a multivariate normal distribution with mean \code{center} and standard deviation \code{sd}.
}
\item{rw.sd}{
- numeric vector; the intensity of the random walk to be applied to parameters.
+ numeric vector with names; the intensity of the random walk to be applied to parameters.
The random walk is only applied to parameters named in \code{pars}.
The algorithm requires that the random walk be nontrivial.
Thus, each element in \code{rw.sd[pars]} must be positive.
Modified: pkg/tests/ou2-mif.Rout.save
===================================================================
--- pkg/tests/ou2-mif.Rout.save 2008-08-25 17:14:21 UTC (rev 50)
+++ pkg/tests/ou2-mif.Rout.save 2008-08-25 21:23:02 UTC (rev 51)
@@ -68,11 +68,11 @@
> mif.fit <- continue(mif.fit,Nmif=70,max.fail=100)
> toc <- Sys.time()
> print(toc-tic)
-Time difference of 1.396902 mins
+Time difference of 1.694885 mins
> cat("PF estimated log likelihood at MIF MLE\n")
PF estimated log likelihood at MIF MLE
> print(pfilter(mif.fit)$loglik,digits=4)
-[1] -421.7
+[1] -422.6
>
> cat("coefficients at truth\n")
coefficients at truth
@@ -83,7 +83,7 @@
MIF MLE
> print(coef(mif.fit,c('x1.0','x2.0','alpha.1','alpha.4')),digits=4)
x1.0 x2.0 alpha.1 alpha.4
- 51.2028 -50.0965 0.9032 0.9738
+ 51.1941 -51.2739 0.8990 0.9736
>
> plot(mif.fit)
> compare.mif(mif.fit)
More information about the pomp-commits
mailing list