ML Workflow CT-AI Practice Questions (Page 5)

Questions 21–25

1. 21expert · hard

   A data scientist has trained a binary classifier and evaluated it on a test set. The model achieves high accuracy, but the business team is concerned about fairness across different demographic groups. The scientist notices that the false positive rate is much higher for one group. What is the most appropriate action?

   • AAnalyze the model's performance using group-specific metrics and consider adjusting the decision threshold for different groups
   • BIgnore the disparity because the overall accuracy is high
   • CRemove the demographic features from the model to ensure fairness
   • DRetrain the model with a different algorithm that is known to be fair
   Reveal Solution Discussion Select an answer first
2. 22application · medium

   A team is training a decision tree model for a classification task. They notice the model performs perfectly on training data but poorly on unseen data. Which approach is most likely to improve generalization?

   • APrune the decision tree to reduce its complexity
   • BUse a larger learning rate
   • CIncrease the depth of the tree to capture more patterns
   • DRemove the validation set to allow more training data
   Reveal Solution Discussion Select an answer first
4. 23foundation · easy

   Which of the following is a typical step in data preparation for ML?

   • ADeploying the model to a production server
   • BAdjusting model parameters to minimize error
   • CSplitting the dataset into training, validation, and test sets
   • DSelecting the ML algorithm to use
   Reveal Solution Discussion Select an answer first
5. 24foundation · easy

   During model training, what does the process of adjusting parameters aim to achieve?

   • AClean the data by removing missing values
   • BMinimize the error between predictions and actual outcomes
   • CIncrease the size of the training dataset
   • DEnsure the model is deployed to production
   Reveal Solution Discussion Select an answer first
6. 25application · medium

   A team has trained a model to predict customer churn. The dataset is imbalanced: only 5% of customers churn. The team wants to evaluate the model's performance. Which metric is most appropriate to focus on?

   • AAccuracy
   • BMean squared error (MSE)
   • CArea under the ROC curve (AUC)
   • DF1 score alone
   Reveal Solution Discussion Select an answer first