Regression
Huber loss
Implemented
regression.huber_lossDefinition
Quadratic for small errors and linear for large errors, controlled by δ.
Formula
Lδ(e) = ½e² if |e| ≤ δ, else δ(|e| − ½δ)
Range: [0, ∞)
Inputs and outputs
- y_true: array of real targets
- y_pred: array of real predictions
- sample_weight: optional
- delta: positive float
Returns: float
Assumptions
No assumptions beyond valid, aligned inputs of the documented types.
Limitations
No metric-specific limitations are documented yet. Interpret the value alongside the task, data, and other metrics.
Python API
import evalsuite as es
es.huber_loss(y_true_r, y_pred_r, delta=1.0)References
- Huber, P. J. (1964). Robust estimation of a location parameter. The Annals of Mathematical Statistics, 35(1), 73–101.