Concept Drift & Fraud Detection — MLA-C01 Practice Question
A representative AWS Machine Learning Engineer Associate (MLA-C01) exam question on Concept Drift & Fraud Detection. Work through it below, then read why each option is right or wrong.
Short answer
The correct answer is B. Concept drift.
Concept drift is when the mapping between features and the target label changes over time. Here the same feature values now correspond to a different label, which is the definition of concept drift.
The Question
An online gaming company's cheat-detection model was trained on player behavior from a year ago. Players have since changed their cheating techniques, so the same behavioral feature values that used to indicate "not cheating" now frequently indicate cheating. Which phenomenon best describes this?
Why B is correct
Concept drift is when the mapping between features and the target label changes over time. Here the same feature values now correspond to a different label, which is the definition of concept drift.
Why the other options are wrong
Data drift describes a change in the input feature distribution itself, not a change in what those inputs mean for the label.
Training-serving skew refers to inconsistencies between how features are computed at training time versus inference time (e.g., different code paths), not a real-world behavior change.
Feature store staleness refers to outdated feature values being served, which is an operational/data-freshness issue, not a shift in the label relationship.
Key idea: Concept Drift & Fraud Detection
Why B is correct: Concept drift is when the mapping between features and the target label changes over time. Here the same feature values now correspond to a different label, which is the definition of concept drift. Why A is wrong: Data drift describes a change in the input feature distribution itself, not a change in what those inputs mean for the label. Why C is wrong: Training-serving skew refers to inconsistencies between how features are computed at training time versus inference time (e.g., different code paths), not a real-world behavior change. Why D is wrong: Feature store staleness refers to outdated feature values being served, which is an operational/data-freshness issue, not a shift in the label relationship. On the MLA-C01 exam, questions in the "ML Solution Monitoring, Maintenance, and Security" domain test whether you can map a scenario's constraints to the right choice. Read the requirement carefully, eliminate options that violate any single constraint, and pick the one that satisfies all of them with the least operational overhead.
Ready to see how you'd score?
Take the free practice quiz and find out which AWS Certified Machine Learning Engineer - Associate domains you need to focus on. No signup required.
Practice 5 similar questions
Same cert, same or adjacent domain. Use these after reviewing the explanation.