Calibration
Calibration slope and intercept
Implemented
calibration.calibration_slopeDefinition
Coefficients of a logistic regression of the outcome on the logit of the predicted probability.
Formula
logit P(y = 1) = a + b · logit(p); ideal a = 0, b = 1
Inputs and outputs
- y_true: array of binary labels {0, 1}
- y_prob: predicted probability of the positive class in [0, 1]
Returns: slope, intercept with confidence intervals
Assumptions
No assumptions beyond valid, aligned inputs of the documented types.
Limitations
- Summarises calibration with two numbers; inspect the calibration curve as well.
Python API
import evalsuite as es
es.calibration_slope(y_true, y_prob)
es.calibration_intercept(y_true, y_prob)References
- Cox, D. R. (1958). Two further applications of a model for binary regression. Biometrika, 45(3/4), 562–565.