krisbitney/StarbucksCapstone

Used Logistic regression with tensor-product natural spline basis and a ridge penalty to explore heterogeneous predictive effects of features using simulated data from a Starbucks marketing campaign

Jupyter Notebook