Wow. Now I am confused. I thought bayesian models were based on priors. So I would just have the model like the below without all the priors?
Y = poisson_log(xB1 + xB2....xBn)
Wow. Now I am confused. I thought bayesian models were based on priors. So I would just have the model like the below without all the priors?
Y = poisson_log(xB1 + xB2....xBn)
The default prior is uniform. It won’t bias the regression but Max mentioned that you’ve got a really large sample . If that’s so then it won’t matter.
If you’re picking and choosing which posterior samples you use, can I suggest picking a higher number of warm up iterations (the ones that don’t get used).
You can suggest anything you want. I want to learn. How about 5000 iterations and 1000 warmup? Is that enough?
Jordan
Sure, just more than you had before to see if it improves your posterior samples.