Hands-On Exercise: Evaluate the Created ML Model CT-AI Practice Questions (Page 2)

Questions 6–10

1. 6application · medium

   A team evaluated a regression model predicting energy consumption. The business requires that the model's predictions be accurate enough for capacity planning. Which metric would be most appropriate to assess the typical magnitude of prediction errors?

   • AMean Absolute Error (MAE)
   • BR-squared (R²)
   • CAccuracy
   • DF1-score
   Reveal Solution Discussion Select an answer first
2. 7application · medium

   A team evaluated a model and must document the results for a compliance audit. Which information is most important to include in the evaluation report?

   • AThe test dataset description, evaluation metrics, and the model's performance against requirements
   • BThe number of training epochs and learning rate
   • CThe exact code used to train the model
   • DThe names of the team members who developed the model
   Reveal Solution Discussion Select an answer first
4. 8expert · hard

   A team developed a model to predict loan defaults. The business requirement is to approve at least 90% of good applicants (high recall for the 'good' class) while keeping the default rate among approved loans below 5% (high precision for the 'good' class). The model currently achieves 92% recall and 96% precision for the 'good' class. Which conclusion is most appropriate?

   • AThe model's performance cannot be assessed without knowing the default rate in the general population.
   • BThe model meets the recall requirement but fails the precision requirement.
   • CThe model meets both requirements, so it is ready for deployment.
   • DThe model fails both requirements because precision and recall are not both 100%.
   Reveal Solution Discussion Select an answer first
5. 9application · medium

   A team evaluated a binary classifier for detecting defective products on an assembly line. The model achieved 95% precision and 60% recall. Which statement best interprets these results?

   • AThe model has a high false positive rate, which is why precision is high.
   • BWhen the model flags a product as defective, it is correct 95% of the time, but it misses 40% of actual defects.
   • CThe model correctly identifies 95% of all defective products, but 40% of flagged products are actually fine.
   • DThe model is unreliable because recall is below 95%.
   Reveal Solution Discussion Select an answer first
6. 10expert · hard

   A medical diagnostics team developed a model to detect a rare disease. The cost of a false negative is extremely high (missing a disease), while false positives lead to additional tests but are less harmful. The model currently achieves 99% recall and 60% precision. The team wants to reduce the number of false positives without significantly increasing false negatives. Which approach is most appropriate?

   • ARaise the decision threshold to increase precision, but monitor recall to ensure it remains acceptable.
   • BLower the decision threshold to increase recall, accepting more false positives.
   • CRetrain the model with more data, but this is unlikely to change the trade-off.
   • DUse a different metric, such as F1-score, to evaluate the model.
   Reveal Solution Discussion Select an answer first