Oversee AI/ML Model Technique(s) (e.g., Algorithm, Selection) PMI-CPMAI Practice Questions (Page 3)

Questions 11–15

1. 11foundation · easy

   Which task is a typical application of unsupervised learning?

   • APredicting whether an email is spam or not using labeled emails
   • BReducing the number of features in a dataset while preserving its essential structure
   • CTeaching a robot to navigate a maze by rewarding it for reaching the goal
   • DForecasting next quarter's sales using historical sales data
   Reveal Solution Discussion Select an answer first
2. 12expert · hard

   A healthcare organization is building a model to predict patient readmission. They have a large dataset with many features, but the model must be interpretable for clinicians and regulatory compliance. The team is debating between a simple logistic regression and a complex gradient boosting model. What is the best decision?

   • AChoose a deep learning model and use SHAP for explanations.
   • BChoose gradient boosting because it will have higher accuracy.
   • CChoose logistic regression because it is fully interpretable.
   • DChoose an unsupervised clustering model to identify risk groups.
   Reveal Solution Discussion Select an answer first
4. 13application · medium

   An e-commerce company wants to predict the likelihood that a user will click on an ad. They have historical data with user features and a binary click/no-click label. The model will be used in real-time bidding where low latency is critical. Which algorithm family is most appropriate?

   • ASupervised classification algorithms like logistic regression or gradient boosting.
   • BUnsupervised clustering to group users by behavior.
   • CReinforcement learning to explore ad placements.
   • DDimensionality reduction with PCA before any modeling.
   Reveal Solution Discussion Select an answer first
5. 14foundation · easy

   What should be included in the algorithm selection criteria documentation to ensure transparency?

   • AThe algorithm's popularity in the tech community
   • BThe personal preferences of the data scientists
   • COnly the final algorithm name without any justification
   • DThe rationale for choosing the algorithm, including trade-offs considered
   Reveal Solution Discussion Select an answer first
6. 15foundation · easy

   When researching candidate algorithms for a classification problem, which factor is most important to consider during the evaluation?

   • AThe algorithm's ability to handle the data characteristics, such as feature types and sample size
   • BThe algorithm's name being familiar to stakeholders
   • CThe algorithm's popularity in online forums
   • DThe algorithm's runtime on a single data point
   Reveal Solution Discussion Select an answer first