In the K = 2 solution, Cluster 1 has far fewer returning-customer orders than Cluster 0 (about 130 versus about 2,860). A marketing manager concludes: “High-value orders come from much less loyal customers, so we should launch a loyalty program aimed at the high-value segment.” What is the best response?
Read DetailsA customer-retention team has limited time and will contact customers predicted to be at risk of churn. Which statements are correct? Select all that apply. Refresh from lessons) Table 5.5: Metrics Cheat Sheet Metric Interpretation Accuracy Overall proportion of correct predictions Precision Of all predicted positives, what proportion is correctly identified Recall (aka sensitivity aka TPR: True Positive Rate) Of actual positives, what proportion is correctly predicted Specificity (aka TNR: True Negative Rate) Of actual negatives, what proportion is correctly predicted FPR aka False Positive Rate = 1- Specificity Of all actual negatives, what proportion is classified incorrectly as positives F1-Score Harmonic mean of precision and recall, balances precision and recall,1: perfect precision and recall, 0: the model is good for nothing ROC-AUC Ability to distinguish between classes across thresholds Log Loss
Read Details