[ENH] Support customized base estimator and predictor

[xuyxu]

resolves #29 #26

Steps

• Implement K-Fold wrapper for base estimators
• Implement customized cascade layer
• Implement set_estimator and set_predictor for the model
• Add unit tests
• Add backward compatibility
• Add documentation and working examples

Code Snippet

from deepforest import CascadeForestClassifier

model = CascadeForestClassifier()

# New Steps
estimator_1, estimator_2 = your_estimator(), your_estimator()
model.set_estimator(estimator=[estimator_1, estimator_2],  # a list of your base estimators
                    n_splits=5,  # the number of folds
                    oob_approx=False,  # whether to use out-of-bag approximation
                    random_state=None)  # random state used for base estimators

model.set_predictor(predictor=your_predictor)  # an instantiated object of your predictor

model.fit(X_train, y_train)
y_pred = model.predict(X_test)

[xuyxu]

Hi @Maryom, this PR implements the feature request on custom base estimators. If you want to use XGBClassifier as the base estimators, below is the example code:

from sklearn.datasets import load_digits
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score

from deepforest import CascadeForestClassifier
from xgboost import XGBClassifier

X, y = load_digits(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=1)
model = CascadeForestClassifier(random_state=1)

# New Steps
estimators = [XGBClassifier() for _ in range(4)]  # 4 base estimators per cascade layer
model.set_estimator(estimators)

model.fit(X_train, y_train)
y_pred = model.predict(X_test)
acc = accuracy_score(y_test, y_pred) * 100
print("\nTesting Accuracy: {:.3f} %".format(acc))

You can find the wheels for installation here .Feel free to comment below if you have any problem or suggestion when using this feature.

[xuyxu]

LGTM.

[Maryom]

Hi @xuyxu I'm really sorry for not being responsive I was sick. Thank you so much for you awesome work 🙏🏼 it is really handy 👌🏼

[kangwenhao]

徐老师，您好！我仿照您的代码段，自定义了xgboost的回归模型，但是一直报错，我有点不知道是为什么？我使用原版深度森林时，没有这样的错误。

[kangwenhao]

下面是正常时的代码

[xuyxu]

Thanks for reporting @kangwenhao, please refer to #67 for details.

[kangwenhao]

Thank you for your help. Good luck!

[kangwenhao]

Teacher Xu, hello.Now I can customize one estimator, but if I want to customize two, there will be a problem，

[xuyxu]

Hi @kangwenhao, please refer to the API Reference for an introduction on set_estimator. For example, the name of the first input parameter should be estimators instead of estimator.