bias_constraint (希望在某个时候,这将使我不再发疯) from fgboost import FGBClassifier as FGBC clf = FGBC() clf.fit(X_train, y_train, s_train) y_prob = clf.predict_proba(X_test)[:,1]