Monitor and Ensure Operational Performance CERTIFIED-DATA-SCIENCE-PRACTITIONER Practice Questions (Page 2)

Questions 6–10

1. 6application · medium

   A model predicts house prices based on features like square footage and location. The team notices that the average square footage of homes in the live data has increased significantly compared to the training data. The model's performance has not yet degraded. What should the team do?

   • ASet up an alert for data drift on the square footage feature to monitor the shift and its impact on performance.
   • BReduce the model's complexity to make it more robust to input changes.
   • CRetrain the model immediately with the new data.
   • DIgnore the shift because the model's performance is still acceptable.
   Reveal Solution Discussion Select an answer first
2. 7foundation · easy

   Which approach is most effective for detecting concept drift in a production model?

   • AComparing input feature distributions to training data
   • BMonitoring the model's prediction accuracy against actual outcomes over time
   • CReviewing the model's hyperparameters
   • DChecking the model's output format
   Reveal Solution Discussion Select an answer first
4. 8application · medium

   A fraud detection model is used to flag suspicious transactions. The cost of a false positive (blocking a legitimate transaction) is high, while missing a fraudulent transaction is also costly but less so. The team wants to monitor the model's performance with a metric that reflects this business trade-off. Which metric should they prioritize?

   • APrecision
   • BF1 score
   • CRecall
   • DAccuracy
   Reveal Solution Discussion Select an answer first
5. 9foundation · easy

   Which Datadog feature is most directly used to visualize model performance metrics over time?

   • ASynthetic tests
   • BDashboards
   • CLog management
   • DAPM traces
   Reveal Solution Discussion Select an answer first
6. 10foundation · easy

   What is the primary purpose of setting up monitoring for a data pipeline?

   • ATo replace the need for model retraining
   • BTo reduce the size of the training dataset
   • CTo track model performance and data quality continuously
   • DTo automate model deployment
   Reveal Solution Discussion Select an answer first