When would you use robust regression (Huber, RANSAC) instead of OLS?
mediumAnswer
- Use robust regression when the data contains outliers that shouldn't dominate the fit.
- OLS minimizes squared error, so a single large-residual point can pull the line dramatically.
- Huber loss is quadratic near zero and linear in the tails, giving less weight to outliers.
- RANSAC fits on random subsets and picks the fit with the most inliers — great when up to ~50% of points are contaminated.
- Quantile regression is another robust option that models a specific quantile.
Check yourself — multiple choice
- Robust regression is faster than OLS
- It downweights or excludes outliers so they don't dominate the fit
- It requires normally distributed features
- It only works for classification
Robust methods (Huber, RANSAC) limit outlier influence — OLS is very sensitive to them.
#linear-regression#outliers
Practise Supervised Learning
215 interview questions in this topic.
Related questions
- Why does OLS have a closed-form solution but logistic regression doesn't?
- What are the classical assumptions of linear regression?
- How do you interpret a coefficient in a multiple linear regression?
- What is VIF and when do you worry about multicollinearity?
- How do you handle strong multicollinearity in a linear model?
- How do you detect and fix heteroscedasticity?