glm(response ~ ., data = your_train_data, family = "binomial")
mdl_churn_vs_both_inter <- glm(has_churned ~ time_since_first_purchase * time_since_last_purchase, data = churn, family = binomial)