New answers tagged


An example can be customer life-time value (CLTV) and churn problems, conducted by Hardie and Fader. In general, the heterogeneity in a cohort is represented by prior distributions over the parameters, which makes it a hierarchical bayesian model. For example, for the churn problem, you could say that everybody in a cohort churns with a probability of $p$. ...

Top 50 recent answers are included