google / lightweight_mmm

LightweightMMM 🦇 is a lightweight Bayesian Marketing Mix Modeling (MMM) library that allows users to easily train MMMs and obtain channel attribution information.
https://lightweight-mmm.readthedocs.io/en/latest/index.html
Apache License 2.0
829 stars 172 forks source link

Divergences and n_eff #302

Open rajat-barve opened 5 months ago

rajat-barve commented 5 months ago

Can someone please explain what is a good number for these metrics (if anything called be "good")? Also, how would I explain it to someone from the business side who is not super strong statistically/technically? Thanks!