cleverly.learners.ParametricFitDiagnostics#
- class cleverly.learners.ParametricFitDiagnostics(learner: str, coef: tuple[float, ...], converged: bool, score_norm: float, iterations: int, reciprocal_condition: float)[source]#
What one parametric fit reports.
- Parameters:
- learner
str The learner class name.
- coef
tupleoffloat The fitted coefficients on the original feature scale, intercept first.
- convergedbool
Whether Newton’s method met its step tolerance. A fit that did not converge raises, so a stored entry always reads
True.- score_norm
float The largest absolute weighted mean score at the fitted coefficients.
- iterations
int The Newton steps taken.
- reciprocal_condition
float The reciprocal condition number of the equilibrated information matrix at the fit.
- learner