'beta_channel' - 'saturation_beta': Ambiguity When Setting Custom Priors for Channels in MMM Configuration

[Denzel124]

Good morning!
I’m reaching out to ask for clarification on the PyMC-Marketing documentation regarding the configuration of custom priors in the MMM model. . I’m currently using the library for a marketing mix modeling project and need to set a prior for the channels in the model.
While reviewing the documentation, I found two possible ways to specify the prior in the my_model_config dictionary:
-As shown in the example notebook, using 'saturation_beta' to refer to the channel prior (https://www.pymc-marketing.io/en/stable/notebooks/mmm/mmm_example.html):

my_model_config = {  
    "saturation_beta": {  
        "dist": "HalfNormal",  
        "kwargs": {"sigma": prior_sigma},  
    },  
}  

-Referring to the prior as 'beta_channel' (https://www.pymc-marketing.io/en/stable/api/generated/pymc_marketing.mmm.mmm.MMM.html)
        my_model_config = { "beta_channel": Prior("LogNormal", mu=np.array([2, 1]), sigma=1), }
Could someone confirm which approach is correct for setting custom priors for the channels? Using 'beta_channel' does not seem to affect the posterior predictions, while changing the 'saturation_beta' prior does produce effects on the posterior predictions. My concern lies in the ambiguity of using 'saturation_beta' as the name for the prior associated with the channels.

[luiztauffer]

it looks like the notebook is outdated: #1190 (comment)

[wd60622]

Custom prior are done with the Prior class. See the latest version of the docs: https://www.pymc-marketing.io/en/latest/notebooks/mmm/mmm_example.html

[wd60622]

"beta_channel" is stale and is not used anymore. Use "saturation_" prefixed variables for the mutliplying effect. That is usually "saturation_beta" but will be "saturation_alpha" for the Menten saturation. I have raised an issue here #1256 to try to catch these earlier