A company uses ML models to predict whether transactions are fraudulent. The company needs to identify as many fraudulent transactions as possible. Which evaluation metric should the company use to…

The correct answer is D. Recall. Recall (also called sensitivity or true positive rate) measures the proportion of actual positive cases (fraudulent transactions) that the model correctly identifies: Recall = TP / (TP + FN). Maximizing recall minimizes false negatives-cases where a fraudulent transaction is…

ML Model Development

Question

A company uses ML models to predict whether transactions are fraudulent. The company needs to identify as many fraudulent transactions as possible. Which evaluation metric should the company use to evaluate the models to meet this requirement?

Options

AF1 score
BArea Under the ROC Curve (AUC)
CPrecision
DRecall

How the community answered

(35 responses)

A
6% (2)
B
11% (4)
C
3% (1)
D
80% (28)

Explanation

Recall (also called sensitivity or true positive rate) measures the proportion of actual positive cases (fraudulent transactions) that the model correctly identifies: Recall = TP / (TP + FN). Maximizing recall minimizes false negatives-cases where a fraudulent transaction is incorrectly classified as legitimate. In fraud detection, missing a fraudulent transaction (false negative) is far more costly than flagging a legitimate one (false positive), so recall is the priority metric. Precision (C) measures how many predicted-fraudulent transactions are actually fraudulent-useful for minimizing false alarms, but not the goal here. F1 score (A) balances precision and recall but does not specifically optimize for catching all fraud. AUC (B) measures overall discriminative ability across thresholds but does not directly express the ability to catch all fraud.

Topics

#Machine Learning Metrics#Recall#Model Evaluation#Fraud Detection

Community Discussion

No community discussion yet for this question.

Full MLA-C01 Practice