Design Use Cases CERTIFIED-DATA-SCIENCE-PRACTITIONER Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   A data scientist has a small dataset and wants to estimate the model's performance reliably without wasting data on a single validation set. Which testing method is most appropriate?

   • AA/B testing
   • BAdversarial testing
   • CCross-validation
   • DHoldout validation
   Reveal Solution Discussion Select an answer first
2. 2foundation · easy

   When designing test cases for a model that classifies customer support tickets, which scenario best represents an edge case?

   • AA ticket with an empty subject line and a very long message body
   • BA ticket written in the same language and format as the training data
   • CA ticket that is a duplicate of a previously classified ticket
   • DA typical ticket with a short subject and a few sentences in the body
   Reveal Solution Discussion Select an answer first
4. 3foundation · easy

   A model that predicts loan approvals is being considered for production. Which situation is a clear trigger for model testing?

   • AThe team wants to select a new database system for storing the model's training data.
   • BThe team wants to create a visual dashboard to show the model's predictions to stakeholders.
   • CThe team wants to refactor the model's code to make it more readable and maintainable.
   • DThe model must be evaluated to ensure it meets regulatory compliance and performs consistently across different demographic groups.
   Reveal Solution Discussion Select an answer first
5. 4expert · hard

   A streaming service wants to test a new recommendation model. They have a large user base and want to measure the impact on user engagement. However, they are concerned about the risk of degrading the experience for a subset of users. Which testing approach best balances the need for reliable results with risk mitigation?

   • ARun an A/B test with a small percentage of traffic and monitor key metrics
   • BDeploy the model to all users immediately
   • CRun a holdout validation on historical data
   • DUse cross-validation on the training data
   Reveal Solution Discussion Select an answer first
6. 5foundation · easy

   Which of the following is a valid testing objective for a model that will be used to approve credit applications?

   • AThe model must be deployed by the end of the quarter.
   • BThe model must be easy to understand for the development team.
   • CThe model must demonstrate fairness by having similar false positive rates across different income groups.
   • DThe model must use the latest machine learning library.
   Reveal Solution Discussion Select an answer first