R Programming Language Certification Machine Learning with R 3

Question 1

What does the confusionMatrix() function from the caret package return for a classification model?

Accepted Answer

A comprehensive summary including sensitivity, specificity, PPV, and NPV

Answer

confusionMatrix() returns the contingency table plus a rich summary of metrics including accuracy, Kappa, sensitivity, specificity, positive/negative predictive values, and more.

Question 2

Which function extracts variable importance scores from a trained caret model object?

Accepted Answer

varImp()

Answer

varImp() is caret's unified function for extracting variable importance from any supported model, normalizing scores to a 0-100 scale for comparability across model types.

Question 3

In R, which function performs Principal Component Analysis (PCA) and is preferred for numerical stability?

Accepted Answer

prcomp()

Answer

prcomp() uses singular value decomposition (SVD) which is numerically more stable than the eigendecomposition used by princomp(), making it the recommended choice for PCA in R.

Question 4

Which R package provides the infrastructure for model stacking (ensemble of multiple trained models)?

Accepted Answer

stacks

Answer

The stacks package (part of tidymodels) implements model stacking by blending predictions from multiple candidate models using a penalized regression meta-learner.

Question 5

What is the main role of the 'tune' package within the tidymodels ecosystem?

Accepted Answer

Hyperparameter tuning and optimization

Answer

The tune package provides functions like tune_grid() and tune_bayes() for searching hyperparameter spaces using cross-validation, integrating with workflows and rsample splits.

R Programming Language Certification Practice Test

R Programming Language Certification Practice Test

R Programming Language Certification Machine Learning with R 3