search

FBartos / RoBSA

RoBSA R package for estimating robust Bayesian survival analyses
8 stars 0 forks source link

how to include random factor(s) in RoBSA function #4

Open lauccy opened 19 hours ago

lauccy commented 19 hours ago

Dear Frantiek,

I am trying to learn and use RoBSA for my data analysis, as common in time-to-event date, my data has a treatment factor and also a random factor. I did not find how to include a random factor in RoBSA function. Please could you guide me how to do a RoBSA model with random factor.

THANKS !!

FBartos commented 16 hours ago

Hi Lauccy,

I'm sorry to disappoint you but the package currently does not allow adding random factors to the model. I might be working on some similar features for other packages of mine, so I will keep it in mind and I might also add this feature in future. I can't promise any timeline unfortunately.

If you are primarily interested in just computing Bayes factors for more complex survival models (rather than model-averaging across parametric survival models), I would recommend using an approximate Bayes factor based on the output from a frequentist model. We illustrated how this can be done in for example this article: https://doi.org/10.1002/sta4.600

Cheers, Frantisek

lauccy commented 7 hours ago

Thank you so much Frantisek! I will try https://doi.org/10.1002/sta4.600.