Free Isaca AAISM Actual Exam Questions - Question 14 Discussion

A imo, cause grouping legit vs threat helps spot false alarms better than just learning history.

I think A sounds like clustering, which is more unsupervised, so probably not the best fit. C focuses on the foundation—learning from labeled examples—which is crucial for reducing false positives. Would D’s data generation really be considered part of supervised learning itself?

C/D? C is key, but D’s about expanding data which also helps reduce false positives.

B tbh doesn’t fit as well since real-time feature engineering sounds more like online learning or adaptive systems, not classic supervised learning. C’s about using labeled data, which is the core of supervised methods.

If it’s just about reducing false positives, isn’t learning from labeled data (C) more direct than grouping (A)?

I think C is the better fit here. Supervised learning mainly depends on labeled historical data to train models, which helps it understand what’s normal and what’s a threat. That foundational step is what reduces false positives, not just pattern grouping like in A. The other options feel less directly tied to how supervised learning works.

Probably C here because supervised models rely on past labeled examples to distinguish threats from normal behavior, which helps reduce false positives by learning known patterns.

It’s A because grouping legitimate activity separately helps cut down false alarms, not just relying on past data like C suggests.

It’s C