Hey everyone, I’m stuck with a CFA in R. When I run my code, I get this warning:
Warning message: In eval(expr, envir, enclos) : Could not compute QR decomposition of Hessian. Optimization probably did not converge.
My model is pretty big with lots of latent variables and covariances. It works fine with fewer variables or when I cut out some of the covariance lines. The calculation takes ages and eats up a ton of memory.
I’m wondering if my 2GB of RAM is the problem here. Or could it be something else I’m missing? Has anyone run into this before? Any tips on how to make it work without buying more RAM?
I’d share my code, but it’s super long. Let me know if you need to see it though. Thanks for any help!
I’ve encountered similar issues with large CFAs in R. While 2GB of RAM is indeed limiting, you might benefit from trying a few alternatives before considering any upgrade. For instance, using a package like lavaan can sometimes handle memory more efficiently. Another approach is to simplify your model by temporarily omitting less critical variables or covariances in order to reduce complexity. You might also check if adjusting the memory settings in R helps at all, though given the hardware, it could be a tough limitation to overcome.
yo maxrock, that sucks man. i had similar probs w/ big models. have u tried using a diff package like OpenMx? it’s pretty good w/ memory stuff. also, maybe try running ur analysis on a cloud service if u can’t upgrade ur RAM. they got way more juice than most laptops. good luck dude!
Hey MaxRock56! 
Oof, memory issues with R can be such a headache, especially when you’re working with big models. I’ve definitely been there before!
Have you tried running your CFA in chunks? Sometimes breaking it down into smaller pieces can help manage the memory load. Also, have you looked into using the lavaan package? It’s pretty memory-efficient for CFAs and might handle your model better.
Just curious - what kind of data are you working with that needs such a complex model? Sounds like an interesting project! 
As for the RAM, 2GB is pretty tight for heavy statistical work. You might be able to squeeze by with some tweaks, but if you’re doing this kind of analysis regularly, upgrading your RAM could save you a lot of headaches in the long run.
Let me know if you want to brainstorm some workarounds. Always happy to put our heads together on tricky R problems!