AI0-001 Machine Learning and Deep Learning Practice Question
A junior ML engineer is asked to evaluate a binary classifier that predicts whether a bank transaction is fraudulent. The model's precision is 0.92 and recall is 0.41. The team wants to improve recall without retraining the model. Which action should the engineer take?
⚠ Common exam trap
The trap here is assuming that changing the evaluation metric (such as switching to accuracy) will change the model's behavior, when metrics only measure performance and do not alter predictions.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Lower the classification threshold for the fraud class.
Recall measures the fraction of actual positives the model captures. When a model has high precision but low recall, it is being too conservative about predicting the positive class. Lowering the decision threshold shifts the operating point along the precision-recall curve toward higher recall, which is exactly what the scenario asks for without retraining the model.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✓
Lower the classification threshold for the fraud class.
Why this is correct
Lowering the decision threshold makes the model more willing to label a transaction as fraudulent, which increases the number of true positives captured and therefore raises recall. Precision will typically drop as more false positives appear, but the scenario explicitly prioritizes improving recall without retraining, so adjusting the threshold is the correct, direct lever.
- ✗
Apply L2 regularization and retrain the classifier.
Why it's wrong here
L2 regularization penalizes large weights and is used to reduce overfitting, not to increase recall on a model that already has high precision but low recall. It also requires retraining, which violates the scenario constraint. Regularization can even reduce sensitivity to rare fraud patterns, potentially lowering recall further rather than improving it.
- ✗
Switch the evaluation metric from F1 score to accuracy.
Why it's wrong here
Accuracy is a poor metric for imbalanced fraud data because a model that predicts 'not fraud' for every transaction can achieve very high accuracy while catching zero fraud. Changing the metric does not change the model's predictions or its recall; it only changes how performance is reported. Therefore it cannot improve recall and is not the right action here.
- ✗
Increase the number of training epochs and re-evaluate.
Why it's wrong here
Increasing epochs retrains the model and does not guarantee higher recall; the scenario states the team wants to improve recall without retraining. Also, additional epochs can worsen overfitting, which might lower generalization performance rather than improve recall on unseen transactions. This action fails the explicit constraint of the scenario and does not directly target the precision-recall trade-off.
About these practice questions
One of 962 original AI0-001 practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official CompTIA exam blueprint
This AI0-001 practice question is part of Courseiva's free CompTIA certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AI0-001 exam.