topic Re: Proc Glimmix: Sparsity Option?? in Statistical Procedures

Proc Glimmix: Sparsity Option??

lkeyes — Fri, 28 Apr 2017 17:51:47 GMT

So, I reciently misspelled an option in the GLIMMIX statment...and it let me know all of the availible options it was expecting to see, and "Sparsity" is one of them.

Does anyone know what that option does?? I haven't found any information on it.

Thanks!

Re: Proc Glimmix: Sparsity Option??

lvm — Sat, 29 Apr 2017 17:13:50 GMT

Procedures often have experimental options that aren't ready for the public, and are not described (they may show up in a production release). These are used by developers and beta testers.

Re: Proc Glimmix: Sparsity Option??

lkeyes — Thu, 04 May 2017 15:25:35 GMT

Thanks @lvm! I figured that's what it was; however, was hoping to hear that it implemented technique simmilar to HPMIXED for large sparse datasets, given a GLMM... /le sigh.

Thanks again!

Re: Proc Glimmix: Sparsity Option??

lvm — Thu, 04 May 2017 17:10:17 GMT

For fitting a GLMM to large data sets, you might find the %hpglimmix macro helpful.

https://www.jstatsoft.org/article/view/v058i08