Regression
Maximum error
Implemented
regression.max_errorDefinition
Largest absolute error: the worst case.
Formula
maxᵢ |ŷᵢ − yᵢ|
Range: [0, ∞)
Inputs and outputs
- y_true: array of real targets
- y_pred: array of real predictions
- sample_weight: optional
Returns: MetricResult
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.max_error(yr, pr)References
- Hyndman RJ, Athanasopoulos G. Forecasting: Principles and Practice. 3rd ed. OTexts; 2021.