How does dataset imbalance influence bias and variance?
Updated May 15, 2026
Short answer
Class imbalance can increase bias toward majority class and increase variance in minority class predictions.
Deep explanation
When datasets are imbalanced, models tend to learn majority class patterns more effectively, leading to biased predictions. Minority classes suffer from poor representation, causing unstable and high-variance predictions. Techniques like resampling, class weighting, and synthetic data generation help mitigate this issue.
Real-world example
Fraud detection systems where fraudulent transactions are rare but critical.
Common mistakes
- Using accuracy as sole metric in imbalanced datasets.
Follow-up questions
- What is SMOTE?
- Which metrics are better for imbalance?