Beta geometric model fit taking forever!

[wfarrar-cil]

Hi guys,

I'm running a CLV analysis per instructions laid out in this article:

https://www.pymc-marketing.io/en/stable/notebooks/clv/clv_quickstart.html#predicting-spend-value-of-customers

I'm using my own dataset and it is in the exact required RFM format and has ~300,000 rows of data.

When I get to this step:

bgm.fit()
bgm.fit_summary()

The loading bar is saying it will take around 78 hours to run the code.

Any ideas why it might be taking so long?

[ricardoV94]

You can try fit(method="map") instead of doing a full posterior fit. How fast does that run for you?