First steps with PyMC and BMA

I am on my first steps with PyMC and BMA, I try to reproduce the model proposed in PyMC and I saw that in the three models mu=0, sigma=10 for alpha. For beta, in the first model mu=0 sigma=10, and for the 2nd and the 3rd model mu=0 sigma=1. Why we choose these values? Sorry for my question and I apologize if I bother you.

Can you provide a link to the notebook you are working with?

https://www.pymc.io/projects/docs/en/v3.11.4/pymc-examples/examples/diagnostics_and_criticism/model_averaging.html

So first, off, the more recent version of this page is here.

Regarding you original question, it does seem like this was perhaps a mistake. @aloctavodia might weigh in (though he wrote the original notebook a long time ago). Alternatively, you (or someone) could check the McElreath book itself (where this example was taken from) to see whether he changes the priors in this way. If not, we would welcome an update. This is exactly the kind of thing we hope to get done at the upcoming PyMC Docathon (detail here).

1 Like

Thank you