Attachment | Size |
---|---|
The principal code used for bivariate ACE model [6] | 36.64 KB |
The code used to carry out batch process [7] | 5.29 KB |
table output [8] | 38.33 KB |
Hi all,
I'm new with OpenMx. I have been trying to use a code obtained from others to carry out a bivariate ACE model. However, I encountered a situation that most of the confidence intervals of the model fit are NA. It seems this topic has been discussed thoroughly on this forum. But I still cannot confirm my problem by myself. I tried switching the optimizer to “SLSQP” but didn't solve the problem. In addition to mxCI, I found there are several alternatives, discussed on this forum, to estimate confidence intervals. Before trying the other options, I want to make sure if it is okay to proceed without confidence intervals. As discussed at (https://openmx.ssri.psu.edu/node/4269), there might be three reasons for the missing value of confidence intervals. I wonder if the fact that the confidence interval is too narrow means the estimate is okay? Or how can I set the limit to verify it? As suggested in the abovementioned discussion, I cannot carry out the analysis. So wish somebody can help me to examine the problem. Thanks a lot.
ps: I attach the code I used to this post.
Below is my mxVersion:
mxVersion()
OpenMx version: 2.17.3 [GIT v2.17.3]
R version: R version 3.6.2 (2019-12-12)
Platform: x86_64-w64-mingw32
Default optimizer: CSOLNP
NPSOL-enabled?: No
OpenMP-enabled?: No
The Results I got:
summary(CholAeFit,verbose=T)
Summary of CholAE
see attachment "ace output.r"