What is missing value imputation in feature engineering?
Updated May 16, 2026
Short answer
Missing value imputation replaces missing data with statistical or predictive estimates.
Deep explanation
Missing values can distort model training. Techniques include mean, median, mode imputation or advanced methods like KNN imputation.
Real-world example
Used in healthcare datasets where patient data is often incomplete.
Common mistakes
- Dropping too many rows leading to data loss.
Follow-up questions
- What is KNN imputation?
- Why is median preferred over mean?