Which method is most appropriate for imputing missing values in a numerical feature with significant outliers?
Key points
- Median is a robust measure of central tendency
- Mean is sensitive to outliers and skewed distributions
- Imputation strategy should match data distribution characteristics
Ready to go further?
