OpenMx Developer Forums

At developers meeting on 3/2 we discussed the following:

• The group discussed several different strategies for Mike Hunter to employ when debugging his LISREL-related code additions to OpenMx. Specifically, it is very difficult to debug R code on Windows because of a lack of gdb support. Mike is going to install a virtual Linux distribution via Virtual Box and attempt to debug through Linux.

When looking at a script, or script snippet, I find it very helpful to see syntax highlighting. I am an addict of highlighting. However, it is difficult to get demo scripts, or script snippets highlighted. One possible solution might be this:

http://alexgorbatchev.com/SyntaxHighlighter/

which currently lacks R support (and its wish list forum has been closed due to spam) but it is customizable. Possibly, building an R highlighter - and particularly one that understands OpenMx functions - would be useful? Or perhaps someone has another solution?

At developers meeting on 2/24 we discussed the following:

• The group discussed Tim Brick's progress with implementing analytic/smart gradients. The gradients save estimations but not iterations in the optimization. Michael Spiegel is going to pursue further performance testing to attempt to determine the possible speedup for various models and scripts.
• The group discussed the status of various grant proposals to fund OpenMx.

I have conducted a large scale numerical simulation study to see small
sample properties of SEM and then have the following question:
I would appreciate it if you could teach me why MLEs in a CFA model are
different between analyses of raw data and covariance matrix data in
OpenMx package of R. Are the optimization methods employed different?
The maximum difference I encountered is 0.084267.

When I compare the results, I adjusted their scales, i.e., multiplied by
sqrt(N/N-1) for factor loadings and by N/N-1 for error variances.

I am not sure whether these are available from NPSOL, but it would be very helpful for diagnostics if the Lagrange multiplier, derivative of constraints, and the augmented (calculated) Hessian matrix could be given in the output slot of a fitted model. My impression is that standard errors are also missing currently when constraints are present.

At developers meeting on 2/17 we discussed the following:

• There is now a GUI for OpenMx installation. The GUI addresses issues with the version of gcc that is provided in Mac OS 10.7.
• More LISREL code is checked in and LISREL objective functions appears close to working.

At developers meeting on 2/10 we discussed the following:

• The implementation of the LISREL objective functions looks good from the back end. Furthermore, the function has been created in a manner that allows additional new types of objective functions to be easily implemented and take advantage of the parallelization improvements in OpenMx 1.2
• The group also discussed the content of several grant proposals for funding for OpenMx.

At developers meeting on 2/3 we discussed the following:

• OpenMx 1.2, whose main feature is parallelization of FIML, will be released this weekend. This will allow the group to point to OpenMx progress in grant proposals and allow it to be tested by real users before the Twins Workshop in Colorado. Updates to the user guide will be seen in future minor release of 1.2 or in a separate OpenMx 1.3 release.

At developers meeting on 1/27 we discussed the following:

• The group discussed a bug in the implementation of the parallel joint and continuous full information likelihood calculation. Tim believes he is narrowing down the cause of the bug and that it will be fixed reasonably soon. Ideally the group would like to release OpenMx 1.2, whose main feature is parallelization of FIML, in the next several weeks. This will allow the group to point to OpenMx progress in grant proposals and allow it to be tested by real users before the Twins Workshop in Colorado.

At the OpenMx Developers Meeting on 1/13/12 we discussed the following:

• The grant application submitted to NIH, that if accepted will help fund the OpenMx project, has been successfully submitted.
• The group discussed the content of a grant application through NSF, that if funded will fund the OpenMx project.