Data Science with Python — Hard
Key points
- Huber loss is a compromise between MSE and MAE
- It is sensitive to small errors and robust to outliers
- The threshold delta determines when to switch from quadratic to linear loss
Ready to go further?
Related questions
