Login

News

OpenMx pre-release r1527

Submitted by mspiegel on

We are pleased to announce the first of our OpenMx pre-release series. As we develop new features for OpenMx, we will periodically make pre-releases available that contain the new functionality. The latest pre-release can be downloaded using the following installation url. You will notice that the pre-releases will always have the version number 999. You can find the true version number by looking at the last four digits of mxVersion() after the "999.0.0-".

models/nightly/BootstrapParallelNightly.R (snowfall test) fails in svn versions 1502 onwards.

Submitted by neale on

If I svn up to any version 1502 or later, and run the BootstrapParallelNightly.R script, it fails with:

> names(submodels) <- omxExtractNames(submodels)
> topModel@submodels <- submodels
>
> modelResults <- mxRun(topModel, silent=TRUE, suppressWarnings=TRUE)
Error in checkForRemoteErrors(val) :
2 nodes produced errors; first error: no slot of name ".estimation" for this object of class "MxModel"

I am running 64bit R: R 2.10.1 GUI 1.31 Leopard build 64-bit (5537).

The error occurs with either snow v. 1.84 or 1.70. Currently have the latter:

OpenMx 1.0.2

Submitted by mspiegel on

This release fixes several minor bugs in the OpenMx 1.0 series. It also represents a transition to R 2.12.x for OS X binary build, and a combination of R 2.11.x and 2.12.x for Windows binary builds. You should be able to use versions of R as early as 2.7 with these binary builds, but you will not have any OpenMx documentation available in R for versions <= 2.9. We are still working on putting together a usable version of npsol for 64-bit windows. Expect to see a binary release of our in-progress 1.1 OpenMx series.

Ordinal error: thresholds in column 'x' are not in sorted order

Submitted by tbates on

#running an ordinal model...
fit = mxRun(model, intervals=F)
Running Full mv ordinal
Error: In model 'Full mv ordinal' the thresholds in column 'integration1' are not in sorted order.

Can this error be made more specific? Does it refer to (mxData as it seems to suggest)?* If so, in which group etc. Or if the thresholds matrix of the matrix pointed to by the "thresholds" parameter of an mxFIMLObjective of a model?

Constraints honoured in models with no objective?

Submitted by tbates on

Hi,
The mini-model below asks that the diag of a matrix "totalVariance" be constrained to 1, but it isn't.
I guess constraints are only heard when optimisation occurs?
t

nVar=3
> test = mxModel("test",
+ mxMatrix("Full", nVar, nVar, free=T, values=0.6, name="totalVariance"),
+ mxMatrix(type="Iden", nrow=nVar, ncol=nVar, name="nVarIdenMatrix"),
+ mxMatrix("Unit", nVar, 1, name="nVarUnit"),
+
+ mxConstraint( diag2vec(totalVariance) == nVarUnit, name="unityVarianceConstraint"),
+
+ mxAlgebra(solve(sqrt(nVarIdenMatrix*totalVariance)), name="iSD")
+ )

mxPath: option to exclude self from all

Submitted by tbates on

If you wish to connect variables each to the other, "all" is currently too powerful, as it also connects each variable to itself

To facilitate scripting (especially the common task of adding drawing covariances between latent variables), it would be nice to have an option to exclude self from all. Given the released function, perhaps the best option would be to add a parameter:

includeSelf = F

Or have I missed an easy trick?

example of current function:
> mxPath(from=latents, arrows=2, free=T, values=1, all=T)
mxPath

Subscribe to
  Login

News