def Agent1PredictorWorker.fit(self, inputs, fx_choice=REG.AIG, **params) [source]
Train the sparse GLM agent on WOE-encoded features.
Executes the full two-stage feature selection pipeline
(stepwise regularization + shadow pruning), fits the
final Logistic Regression, and computes validation
diagnostics.
Parameters
inputs : Agent1Input
Training configuration including train_set and
woe_features.
woe_features.
fx_choice : str
The fitting strategy identifier. Defaults to REG.AIG.
**params : dict
Runtime parameters passed to the internal fitter
(glm_alpha, glm_tol, shadow_quantile, cross_val_folds).
(glm_alpha, glm_tol, shadow_quantile, cross_val_folds).
Returns
Agent1Predictor
The fitted agent instance.