Objective 4.3 Evaluate Models CERTIFIED-DATA-SCIENCE-PRACTITIONER Practice Questions (Page 3)

Questions 11–15

1. 11foundation · easy

   What is the most reliable basis for selecting the best-performing model among candidates?

   • APerformance on a held-out validation or test set
   • BThe model with the most parameters
   • CPerformance on the training set
   • DThe model that trains the fastest
   Reveal Solution Discussion Select an answer first
2. 12foundation · easy

   When comparing two classification models, which approach is most appropriate?

   • ACompare their training time and choose the faster one
   • BCompare the number of layers in each model
   • CCompare their performance on the same validation or test set using relevant metrics
   • DCompare their accuracy on the training set only
   Reveal Solution Discussion Select an answer first
4. 13application · medium

   A team is evaluating a binary classifier for a rare disease with only 1% prevalence. The model achieves 99% accuracy on the test set. The team is concerned that accuracy is not a meaningful metric here. Which additional metric would provide the most insight into the model's ability to detect the disease?

   • ASilhouette score
   • BMean absolute error
   • CRecall (sensitivity)
   • DR-squared
   Reveal Solution Discussion Select an answer first
5. 14application · medium

   A data scientist builds a classifier to predict customer churn. On a test set of 200 customers, the model predicts 50 customers will churn, but only 40 of those actually churn. Additionally, the model fails to predict 10 customers who actually churn. What is the precision of the model?

   • A0.50
   • B0.67
   • C0.20
   • D0.80
   Reveal Solution Discussion Select an answer first
6. 15expert · hard

   A team uses Kubeflow to orchestrate a training pipeline. They need to compare the performance of two different model architectures on the same dataset. They have created two pipeline runs. What is the most efficient way to compare the results?

   • AManually read the logs from each run.
   • BExport metrics to a spreadsheet and compare.
   • CUse Kubeflow Pipelines' run comparison feature to view metrics side by side.
   • DRe-run the pipeline with both architectures in a single run.
   Reveal Solution Discussion Select an answer first