Extended Classifier Methods

model = Classifier("data.csv", target="label").train()

model.cross_validate(cv=5)           # test on 5 splits
model.confusion_matrix()             # breakdown of correct/incorrect predictions
model.roc_curve()                    # ROC curve + AUC (binary only)
model.precision_recall()             # precision-recall curve
model.log_loss()                     # log loss
model.cohen_kappa()                  # accuracy adjusted for chance
model.calibration_curve()            # are confidence scores trustworthy?

model.tune(param_grid={"n_estimators": [50, 100, 200]})
model.tune_random(n_iter=50)

model.learning_curve()
model.validation_curve()
model.permutation_importance()
model.partial_dependence("age")
model.decision_boundary("age", "income")
model.calibrate(method="isotonic")

model.save("my_model.pkl")
model = Classifier.load("my_model.pkl")