Oversee AI/ML Model Technique(s) (e.g., Algorithm, Selection) PMI-CPMAI Practice Questions (Page 1)

Questions 1–5

1. 1application · medium

   A logistics company has high-dimensional sensor data from delivery vehicles. They want to reduce the number of features for visualization and to improve the efficiency of downstream supervised models, but they do not have labels for the sensor data. Which technique should be used?

   • APrincipal component analysis (PCA)
   • BQ-learning
   • CLogistic regression
   • DDecision tree classification
   Reveal Solution Discussion Select an answer first
2. 2expert · hard

   A project manager is coordinating a team that is split between using a traditional machine learning model and a deep learning model for a computer vision task. The deep learning model promises higher accuracy but requires significantly more computational resources and longer training time. The project has a tight deadline and limited GPU budget. What should the project manager do?

   • AImmediately approve the deep learning model to maximize accuracy.
   • BLet the data scientists decide without project manager involvement.
   • CChoose the traditional model to save time and resources.
   • DFacilitate a trade-off analysis to compare accuracy gains against resource and time costs.
   Reveal Solution Discussion Select an answer first
4. 3application · medium

   A project manager is overseeing a cross-functional team building a recommendation system. The data scientists propose a collaborative filtering approach, while the engineering team prefers a content-based approach due to infrastructure constraints. What should the project manager do to ensure alignment?

   • ADelay the decision until after development starts.
   • BChoose the engineering team's approach to avoid infrastructure changes.
   • CSupport the data scientists' proposal because they are the experts.
   • DFacilitate a discussion to evaluate both approaches against project goals and constraints.
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A manufacturing company wants to predict equipment failure. They have a large labeled dataset and need high accuracy to minimize downtime, but they also need to explain the model's decisions to maintenance engineers. Which model choice best balances accuracy and interpretability?

   • AAn unsupervised anomaly detection model.
   • BA gradient boosting model with feature importance scores.
   • CA simple linear regression model.
   • DA deep neural network with high accuracy but no explanation.
   Reveal Solution Discussion Select an answer first
6. 5expert · hard

   A financial services firm is developing a model to detect fraudulent transactions. The data is highly imbalanced (few fraud cases), and the cost of false negatives is very high. The team is considering several algorithms. Which approach is most appropriate given the need to minimize false negatives?

   • AUse a simple logistic regression without handling imbalance.
   • BUse a deep learning model with default settings.
   • CUse a supervised algorithm with class weighting or resampling to handle imbalance.
   • DUse an unsupervised anomaly detection algorithm without any adjustment.
   Reveal Solution Discussion Select an answer first