Regression
Explained variance
Implemented
regression.explained_varianceDefinition
Proportion of target variance explained, ignoring systematic bias (unlike R²).
Formula
1 − Var(y − ŷ) / Var(y)
Range: (−∞, 1]
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.explained_variance(yr, pr)References
- Draper NR, Smith H. Applied Regression Analysis. 3rd ed. Wiley; 1998.