Regression
Median absolute error
Implemented
regression.median_absolute_errorDefinition
Median of absolute errors; robust to outliers.
Formula
MedAE = median(|yᵢ − ŷᵢ|)
Range: [0, ∞)
Inputs and outputs
- y_true: array of real targets
- y_pred: array of real predictions
- sample_weight: optional
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.median_absolute_error(y_true_r, y_pred_r)References
Primary references for this entry have not been added yet.