AI0-001 Machine Learning and Deep Learning Practice Question
A fraud-detection team at a bank trains a gradient-boosted tree model on two years of transaction data. Only 0.4% of transactions are fraudulent. The model achieves 99.7% accuracy but flags almost no fraud. Which approach best addresses the underlying problem with how the model is being trained and evaluated?
⚠ Common exam trap
The trap here is assuming that a 99.7% accuracy score means the model is performing well, when in fact it is simply predicting the majority class.
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
✓
Optimize the model using precision-recall AUC and apply class weighting or resampling to the fraudulent class.
Accuracy is a poor metric when one class is extremely rare, because a trivial majority-class predictor scores deceptively high. The real fix is twofold: train with class weighting or resampling so the learner sees the minority class, and evaluate with precision-recall AUC, which reflects performance on the fraud class rather than being swamped by true negatives.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Remove outliers using a z-score filter and standardize all numeric features before retraining the model.
Why it's wrong here
Preprocessing outliers and scaling features can stabilize training, but it does nothing about the extreme class imbalance that causes the model to ignore fraud. The model already reaches 99.7% accuracy, so feature scaling is not the bottleneck. This distractor sounds like standard ML hygiene but leaves the core metric and class-weight problem untouched.
- ✗
Increase the number of boosting rounds and lower the learning rate until training accuracy reaches 100%.
Why it's wrong here
Adding rounds and shrinking the learning rate pushes the model to fit the training set more closely, but the problem is not insufficient fitting capacity. With a 0.4% positive rate, the loss is still dominated by the majority class, so the trees keep modeling legitimate transactions. This also risks overfitting and makes the reported accuracy look even better while fraud recall stays near zero.
- ✓
Optimize the model using precision-recall AUC and apply class weighting or resampling to the fraudulent class.
Why this is correct
Accuracy is misleading on a 0.4% positive class because a model that predicts 'not fraud' every time scores 99.6% accuracy while catching nothing. Precision-recall AUC focuses on the minority class, and class weighting or resampling forces the learner to pay attention to fraudulent examples. Together they fix both the training signal and the evaluation metric, which is exactly the failure observed.
- ✗
Switch the evaluation metric to root mean squared error and report it alongside accuracy for each boosting round.
Why it's wrong here
RMSE is a regression metric and is not appropriate for a binary fraud classification task. Even if it were computed on predicted probabilities, it would still be dominated by the overwhelming majority of legitimate transactions, so it would not reveal the near-zero fraud recall. Changing the metric without addressing class imbalance leaves the actual problem unsolved.
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.