To declare stacking classifier, we can use:
StackingClassifier(estimators=[
("NB", MultinomialNB()),
("SVM", LinearSVC()),
("LR", LogisticRegression()
]
I wonder if there is a way to declare multiple stacking layers, i.e.:
StackingClassifier(estimators=[
StackingClassifier(estimators=[
("NB", MultinomialNB()),
("SVM", LinearSVC()),
("LR", LogisticRegression()),
]),
VotingClassifier(estimators=[
("NB", MultinomialNB()),
("SVM", LinearSVC()),
("LR", LogisticRegression()),
])
],
final_estimator = GradientBoostingClassifier(), n_jobs=-1
)
question from:
https://stackoverflow.com/questions/65880022/is-there-a-way-to-stack-multiple-layers 与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…