Colt Allen
Colt Allen
## Description of your problem or feature request [I'm currently rewriting the Lifetimes library to run on PyMC](https://github.com/ColtAllen/btyd), but the likelihood function of one of the models requires a [Gaussian...
This PR resurrects https://github.com/pymc-labs/pymc-marketing/pull/188 to partially address https://github.com/pymc-labs/pymc-marketing/issues/176. The Beta-Geometic/Beta Binomial model is a CLV model for discrete, non-contractual use cases - a good example would be sporting events. This...
[RFM Segmentation](https://www.optimove.com/resources/learning-center/rfm-segmentation) is a way to segment customers on their past purchasing behavior. This is a complementary technique to CLV/BTYD modeling as customers in different segments benefit from different marketing...
Now that we have a [Dockerfile for development](https://github.com/pymc-labs/pymc-marketing/tree/main/scripts/docker), I want to get everyone's thoughts on adding a `.devcontainer/devcontainer.json` file for GitHub Codespaces. This would enable anyone signed into GitHub to...
While working on https://github.com/pymc-labs/pymc-marketing/pull/177, I found quite a few things in the literature that can be added to this model, but I'd rather do so in separate PRs: - [...
`lifetimes` contains a variety of plotting functions for model evaluation: - [ ] `plot_period_transactions` - [ ] `plot_calibration_purchases_vs_holdout_purchases` - [x] `plot_frequency_recency_matrix` - [x] `plot_probability_alive_matrix` - [ ] `plot_expected_repeat_purchases` - [...
If a fitted `GammaGammaModel` has a `q` parameter less than 1, all predicted spend values will be negative. This usually happens when the model is fit to heavy-tailed monetary value...
`expected_customer_lifetime_value` is essentially a FOR loop across the specified time range, and it is initialized with an array of zeroes. However, instead of simply creating an array of zeros, model...
The idea of adding discount rates that vary by customer to CLV estimations came up in a recent modeling discussion, and I'm curious to get everyone's thoughts.
`rfm_summary` is the primary data preprocessing step for CLV modeling. It currently returns a pandas dataframe, but [recent changes to the `clv` API](https://github.com/pymc-labs/pymc-marketing/issues/527) are more conducive to Xarrays for parameter...