search

gertvv / gemtc

GeMTC R package: model generation for network meta-analysis
GNU General Public License v3.0
43 stars 25 forks source link

Support for Standardized Mean Difference input? #40

Open joelkuiper opened 8 years ago

joelkuiper commented 8 years ago

I assume (with your new job and all, congratz!) that you won't have a tremendous amount of time implementing this. But maybe you know some literature so I can give it a stab and send a PR when/if done?

gertvv commented 8 years ago

This would be great actually, its on our roadmap, at least I think you mean input? It should be pretty straightforward. My preference would be for it to be a new likelihood, rather than having it be done through pre-processing of the data. Possibly "StandardizedNormal" works as the likelihood name?

joelkuiper commented 8 years ago

Yes I meant input! changed the title too. Potentially I have some code for doing it by pre-processing the data, I guess doing it with a new likelihood would probably entail lifting that code and putting it in the the dedicated likelihood code, or would you rather have JAGS deal with it?

gertvv commented 7 years ago

I have started a branch for this, some work remains (e.g. starting values), but it seems to work.

https://github.com/gertvv/gemtc/tree/feature/smd

If you have a good smallish example dataset, that would help. The only example I have is huge and confidential, so not great for a validation / regression test.