AWS Certified Machine Learning Engineer Associate · Free Practice Question Medium
Question 39
A company is developing an AI solution that needs to reduce both underfitting and overfitting to ensure a balanced model. Which combination of strategies can the data scientist apply to mitigate bias and variance in the model? (Choose Two)
-
A
Use early stopping to prevent the model from overfitting during training.
-
B
Use cross-validation to assess model performance on multiple data subsets.
-
C
Ignore high variance to focus on reducing bias.
-
D
Limit the dataset size to reduce noise in the model.
-
E
Increase the model complexity by adding more parameters.
Reveal correct answers
Correct answers: A, B
A.
Early stopping helps prevent overfitting by halting the training process when the performance on the validation data begins to degrade, indicating the model is starting to memorize noise rather than learning useful patterns.
B.
Cross-validation helps in evaluating the model’s performance on different subsets of data, ensuring that it generalizes better and is not overly fitted to a specific portion of the dataset.
C.
High variance should not be ignored, as it leads to overfitting. Both bias and variance need to be managed in a balanced way (bias-variance trade-off).
D.
Reducing the dataset size does not solve the problem of noise but may limit the model's ability to capture all relevant patterns in the data.
E.
Increasing model complexity by adding more parameters could lead to overfitting, not reducing it.
Discussion
Think the marked answer is wrong, or have a better explanation? Share it below — comments appear after review.
