StackedML
Practice
Labs
Questions
Models
Pricing
Sign in
Questions
/
Supervised Learning
/
Data Issues
/
Imbalanced data handling
← Previous
Next →
797.
Undersampling and Its Drawback
medium
What is undersampling and what is its main drawback?
A
Applying a lower classification threshold to the majority class to reduce the number of false positives
B
Removing outliers from both classes to create a cleaner training set before model training begins; it reduces model robustness to edge cases
C
Reducing the majority class to match the minority class size; it risks discarding potentially informative majority class samples
D
Scaling down feature values in the majority class to reduce its influence on distance-based metrics
Sign in to verify your answer
← Back to Questions