Foundational Concepts DY0-001 Practice Questions (Page 1)

Questions 1–5

1. 1application · medium

   A data scientist is training a gradient-boosted tree model on a dataset with 50,000 rows and 200 features. The training loss decreases steadily, but the validation loss begins to increase after 40 boosting rounds. The team needs a model that generalizes well for next month's deployment. Which combination of actions should the data scientist take?

   • ARemove the validation set and train for more rounds on the full dataset
   • BIncrease the learning rate and increase the maximum tree depth
   • CAdd L1 regularization and enable early stopping on the validation set
   • DDecrease the learning rate and increase the number of boosting rounds
   Reveal Solution Discussion Select an answer first
2. 2application · easy

   A team is training a large logistic regression model with hundreds of features. The model performs very well on the training data but poorly on the validation data. The team suspects overfitting. Which regularization technique is most appropriate to reduce the model's complexity and improve generalization?

   • AL2 regularization (Ridge)
   • BIncrease the number of training epochs
   • CRemove the regularization term entirely
   • DIncrease the learning rate
   Reveal Solution Discussion Select an answer first
4. 3foundation · easy

   Which regularization technique adds a penalty equal to the sum of the absolute values of the model coefficients?

   • AEarly stopping
   • BDropout
   • CL2 regularization (Ridge)
   • DL1 regularization (Lasso)
   Reveal Solution Discussion Select an answer first
5. 4foundation · easy

   Which loss function is commonly used for regression tasks and is less sensitive to outliers than Mean Squared Error?

   • ABinary Cross-Entropy
   • BCategorical Cross-Entropy
   • CMean Absolute Error (MAE)
   • DHinge Loss
   Reveal Solution Discussion Select an answer first
6. 5expert · hard

   A team is building a model to predict equipment failure in a manufacturing plant. They have a dataset with 10,000 records and 50 features. The data is imbalanced, with failures occurring in only 5% of cases. They have trained a single decision tree, which has high variance. They are considering using a random forest or a gradient boosting machine (GBM). Which statement correctly compares these two ensemble methods in this context?

   • AA random forest will always outperform a GBM on imbalanced datasets.
   • BBoth random forest and GBM are equally sensitive to outliers and require the same amount of hyperparameter tuning.
   • CA random forest is generally more robust to outliers and requires less hyperparameter tuning than a GBM.
   • DA GBM is generally more robust to outliers and requires less hyperparameter tuning than a random forest.
   Reveal Solution Discussion Select an answer first