Lesson 4 Flashcards
1
Q
What is the difference between Confidence and Prediction Intervals?
A
- Confidence intervals ignore the errors
- Prediction intervals don’t and are wider
2
Q
How does K-fold cross-validation improve over simple split-sample validation?
A
It divides the data into K subsets and rotates the testing fold across K iterations, averaging performance across all folds.
3
Q
What is a key limitation of two-way splits in model evaluation?
A
They can underestimate prediction error by using the same data for model selection and error reporting.
4
Q
What is the output of K-fold cross-validation for each model?
A
An average MSE value across K validation sets.
5
Q
A
