Code QuizIntermediate Soft Voting Without predict_proba
Soft voting fails when a base estimator can't produce probabilities.
Codepython
from sklearn.ensemble import VotingClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
clf = VotingClassifier(
estimators=[('lr', LogisticRegression()), ('svc', SVC())],
voting='soft'
)
clf.fit(X_train, y_train)This soft-voting ensemble raises an error at fit/predict time. What is the bug?