如何将 sklearn 的 TransformedTargetRegressor 与自定义数据转换器一起使用？

Question

我想使用sklearn.compose.TransformedTargetRegressor，如本答案所示。然而，转换器是自定义的，我遇到了一个错误。

在这个最小的例子中，目标值应该乘以10，然后在预测时再次除以10。(在我的实际应用程序中，目标值必须从非数字格式转换为数字格式。(

import numpy as np
import sklearn
from sklearn.compose import TransformedTargetRegressor
from sklearn.linear_model import LinearRegression
class MyTransform(sklearn.base.TransformerMixin):
def fit(self, *_, **__):
return self
def transform(self, X):
return np.array(X)*10
def inverse_transform(self, X):
return np.array(X)/10

def MyLinearRegression():
return TransformedTargetRegressor(
regressor=LinearRegression(),
transformer=MyTransform()
)

if __name__ == '__main__':
model = MyLinearRegression()
model.fit(X=[[1], [2], [3]], y=[1, 2, 3]) # raises TypeError

这引发了：

Traceback (most recent call last):
File "<input>", line 1, in <module>
File "C:Program FilesJetBrainsPyCharm Community Edition 2019.3.3pluginspython-cehelperspydev_pydev_bundlepydev_umd.py", line 197, in runfile
pydev_imports.execfile(filename, global_vars, local_vars)  # execute the script
File "C:Program FilesJetBrainsPyCharm Community Edition 2019.3.3pluginspython-cehelperspydev_pydev_imps_pydev_execfile.py", line 18, in execfile
exec(compile(contents+"n", file, 'exec'), glob, loc)
File "C:/Users/me/.PyCharmCE2019.3/config/scratches/scratch.py", line 26, in <module>
model.fit(X=[[1], [2], [3]], y=[1, 2, 3]) # raises TypeError
File "C:Usersme.virtualenvsproject--3333Ox_libsite-packagessklearncompose_target.py", line 185, in fit
self._fit_transformer(y_2d)
File "C:Usersme.virtualenvsproject--3333Ox_libsite-packagessklearncompose_target.py", line 127, in _fit_transformer
self.transformer_ = clone(self.transformer)
File "C:Usersme.virtualenvsproject--3333Ox_libsite-packagessklearnbase.py", line 64, in clone
raise TypeError("Cannot clone object '%s' (type %s): "
TypeError: Cannot clone object '<__main__.MyTransform object at 0x000001653B2FA9A0>' (type <class '__main__.MyTransform'>): it does not seem to be a scikit-learn estimator as it does not implement a 'get_params' methods.

Answer 1

您只需要从sklearn.base.BaseEstimator以及transformerixin继承即可：(。类型错误显示：

它似乎不是一个scikit学习估计器

所以你只需要把它变成一个：D。下面的代码应该可以工作。

import numpy as np
import sklearn
from sklearn.compose import TransformedTargetRegressor
from sklearn.linear_model import LinearRegression
class MyTransform(sklearn.base.BaseEstimator, sklearn.base.TransformerMixin):
def fit(self, *_, **__):
return self
def transform(self, X):
return np.array(X)*10
def inverse_transform(self, X):
return np.array(X)/10

def MyLinearRegression():
return TransformedTargetRegressor(
regressor=LinearRegression(),
transformer=MyTransform()
)

model = MyLinearRegression()
model.fit(X=[[1], [2], [3]], y=[1, 2, 3])