New Feature Proposals
AdminMike
Joined: 10/07/2009
New Feature Proposals
This forum is for developers to proposal schemes for the implementation of new features. If you post a new topic, please note that you will be the goto person for implementing the new feature. Otherwise, you are probably looking for the OpenMx wishlist forums. Comments from all users are always welcome on any existing proposal.
- Read more about New Feature Proposals
- Log in or register to post comments
yaning
Joined: 04/12/2019
gene based gene environment interaction in twin studies
Hello, everyone
I’m going to explore gene environment interaction based gene in twin studies. I’ve searched for literatures, and found general/generalized linear mixed model framework of ACE model may be good choices(https://www.ncbi.nlm.nih.gov/pubmed/28953253).
Unfortunately, I'm terrible at math and basic theory. And I cannot find the relevant code.
So I want to ask if anyone has done any research that can give some help.
Thank you!
tbates
Joined: 07/31/2009
allow mxOption to get as well as set
currently
mxOption(m1, key="Number of Threads")
and
mxOption(m1)
Both fail (no value, no key errors, respectively)
It would be nice if asking for a key returned its value, and asking about a model returned the currently active options
- Read more about allow mxOption to get as well as set
- 2 comments
- Log in or register to post comments
tbates
Joined: 07/31/2009
diag()
I'd really like diag implemented in Openmx - just so that the same code works inside and outside of OpenMx and to ease the learning transition. But, If we aren’t going to support the standard r implementation of diag(), can we implement
diag <- function(x=1, nrow, ncol) {
message("We’re a bit worried that diag can be ambiguous, so we’ve broken it into multiple new functions...")
# then either say
stop("You need to choose among list up our replacement functions to handle everything that diag() does...")
- Read more about diag()
- Log in or register to post comments
Ryne
Joined: 07/31/2009
Many many submodels (how S4 builds objects and a request to run lists of models)
So I'm following up on some of my factor score work and happened across and old question I had regarding holder or parent models for embarassingly many submodels.
I'm trying to run the same model on 500 (or some other large n) datasets with the independence flag set to TRUE. I'm doing this by creating one model per dataset, building a list of mxModels, then putting this list into a single mxModel for optimization. However, I'm spending most of my time the fourth line of this code:
singScore <- transformFactorScores(spRes, 1, "mu", "sigma", "epsilon")
neale
Joined: 07/31/2009
Cholesky Matrix Function
Classic Mx had a \chol function to obtain L for symmetric positive (semi-)definite A in A = L * L'. I figure that the BLAS function
dpbtrf could be used: SUBROUTINE DPBTRF( UPLO, N, KD, AB, LDAB, INFO )
and that it wouldn't be too hard to add this to the list of omxAlgebraFunctions.c - but that it would be harder for me than for say Tim Brick to do this. I made a start but have been hauled off to do other frantically urgent things. I suspect we need to tell the algebra cruncher routine the priority of this operation also.
- Read more about Cholesky Matrix Function
- 4 comments
- Log in or register to post comments
neale
Joined: 07/31/2009
Improve performance of R front end of OpenMx
So I have this model and it takes 800 seconds to compile in the front end. Yes, it has some 81 submodels in it, but still... Classic Mx had finished running the script in less time whereas Open is still going after 9hrs (preliminary results only, it is possible that model is mis-specified).
This thread caught my eye: http://www.r-bloggers.com/the-new-r-compiler-package-in-r-2-13-0-some-first-experiments/
Is it worth trying to compile some bits of the OpenMx frontend to improve performance there?
tbates
Joined: 07/31/2009
mxCompare: make comparison optional
This is a minor proposal, but I'd like this to work:
> mxCompare2(fit1)
base comparison ep minus2LL df AIC diffLL diffdf p
1 ACE_Non_ScSL_Cor_a
Currently it would error: Comparison model missing.
- Read more about mxCompare: make comparison optional
- 1 comment
- Log in or register to post comments
Ryne
Joined: 07/31/2009
Correct behavior for 'all=TRUE' for symmetric paths.
Motivated by http://openmx.psyc.virginia.edu/thread/996, Tim (Brick) and I have been talking about how mxPath behaves when arrows=2 and all=TRUE. The current behavior creates all possible paths, save for when excludeself is set to TRUE, in which case the paths from a variable to itself are created then ignored as best as I can figure (I wasn't there/don't remember the meeting for this spec).
tbates
Joined: 07/31/2009
AIC(fit1)
would it be worth getting this to work for Mx as an accessor function for AIC, with the benefit of being consilient with other functions in R?
AIC(fit1)
Error in UseMethod("logLik") :
no applicable method for 'logLik' applied to an object of class "MxModel"
- Read more about AIC(fit1)
- 3 comments
- Log in or register to post comments