Free Isaca AAIA Actual Exam Questions - Question 9 Discussion
Model Card – Electrical Grid Predictive Maintenance Model
Model Information:
Description: AI model designed to predict maintenance needs for electrical grid components, reduce
unplanned downtime, and improve grid reliability.
Inputs: Real-time sensor data, historical maintenance records, and operational logs.
Outputs: Maintenance needs predictions for 60 & 90 days.Evaluation:
Approach: Cross-validation and validation of accuracy, precision, and recall.
Results: Accuracy 72%; Precision 60%; Recall 95%; F1 76%
D, since F1 score really reflects the balance between precision and recall, not just true positives.
Not B, uptime isn’t mentioned or inferred from accuracy or recall here.
Not A, because 95% is recall, which means it catches most true cases but doesn’t mean the model is correct 95% overall. D fits better since F1 balances precision and recall.
This one’s tricky, but I’d go with A because recall being 95% means the model catches almost all true maintenance needs, which is what “correctly predicts” hints at. The 72% accuracy and 60% precision don’t directly support D as strongly since F1 combines both, but high recall is more about correct positives. Plus, options B and C don’t seem related to the model’s actual output. So A fits best with the idea of correctly predicting maintenance needs most of the time.
D vs A? F1 score is about balance, not just accuracy like 95% recall suggests.
Option D looks best since F1 score combines precision and recall, showing how well the model identifies true maintenance needs overall. The others don’t quite match the stats given.