I want Identifying Pre Selection in PyMC3 Bayesian Modeling

1

Hello PyMC3 community :slightly_smiling_face:

I am new to Bayesian modeling and have been using PyMC3 as my go-to tool. While I have made some progress with the fundamentals, I am still having trouble with selecting the right priors for my models.

What categories of priors are accessible in PyMC3? Do you follow to any best practices or suggestions while choosing priors. How can one balance non informative priors that let the data speak for itself with informative priors that take expert knowledge into account?

I have read some tutorials & also checked this- https://discourse.pymc.io/t/initial-release-of-pymc-marketing-bayesian-mmm-and-clv-modeling-muleSoft/11829
but sill want your guidance.
Could anyone here provide some guidance or resources. Any insights or personal experiences you can share would be greatly appreciated.

Thank you all in advance for your help. I want to diving deeper into Bayesian modeling with PyMC3 with your guidance.

Best regards,
Leila

2

Given where you’re coming from, I’d suggest Richard McElreath’s book Statistical Rethinking.

Remember that when someone says “let the data speak for themselves,” they are urging you to look at one experiment in isolation from all of scientific history. Also remember when getting hung up on priors that the likelihood is a much much bigger subjective choice in your modeling!

2 Likes