Objective 7.1 Report Findings CERTIFIED-DATA-SCIENCE-PRACTITIONER Practice Questions (Page 10)

Questions 46–50

1. 46foundation · easy

   What is the primary purpose of the ELI5 library in machine learning?

   • ATo deploy models to production
   • BTo visualize data distributions
   • CTo train deep learning models
   • DTo debug and explain machine learning models
   Reveal Solution Discussion Select an answer first
2. 47foundation · easy

   Which outcome is a typical result of conducting user testing on a data science deliverable?

   • AA list of usability issues and suggestions for improving the user interface
   • BA data dictionary documenting the meaning of each field in the dataset
   • CA statistical report of the model's precision and recall on the test set
   • DA deployment plan for moving the model to a production environment
   Reveal Solution Discussion Select an answer first
4. 48application · medium

   A healthcare startup uses a deep neural network to predict patient readmission risk. Regulators require an explanation for each individual patient's risk score. The team cannot replace the neural network because it outperforms simpler models. Which approach best satisfies the regulatory requirement?

   • AReport only the global feature importance from the neural network's weights.
   • BReplace the neural network with a logistic regression model to gain inherent interpretability.
   • CProvide the patient's raw risk score without any feature-level explanation.
   • DUse LIME to generate a local surrogate model that approximates the neural network's decision for each patient.
   Reveal Solution Discussion Select an answer first
5. 49foundation · easy

   Which of the following best describes the scope of a basic web application used as a proof-of-concept for a machine learning model?

   • AA mobile application with push notifications and offline support
   • BA comprehensive analytics dashboard with historical data and trend analysis
   • CA fully featured application with user authentication, database integration, and load balancing
   • DA minimal application that accepts input, runs the model, and displays the prediction
   Reveal Solution Discussion Select an answer first
6. 50application · medium

   A data science team is building a model to predict employee attrition. The HR department will use the model to identify at-risk employees, while employees themselves may see the outputs. The team needs to present findings. Which stakeholder consideration is most important for the presentation?

   • AFocus exclusively on the model's accuracy and ignore the human impact.
   • BPresent only the technical metrics (AUC, precision, recall) to demonstrate model quality.
   • CTailor the presentation to HR's need for actionable insights while being mindful that employees may see the results and could be affected by them.
   • DShare the raw data with all stakeholders to ensure transparency.
   Reveal Solution Discussion Select an answer first