All,
I've been working on Machine Learning based Membership churn/retention Modeling for the past month or so. I've got a prototype model working with a fairly consistent ROC AUC 0.82 - 0.85, and a 0.75 - 0.76 accuracy using ML methods like Random Forest, Gradient Boosted Trees, and XGBoost.
I'm interested in talking shop and exchanging notes with folks who are interested in doing the same kinds of thing or have produced similar models.
I'm particularly interested in the tools you are using, features you are using. Challenges you may be having linking data from the time of the events. (That is "AS OF" reporting)