如何在机器学习中的数值和分类特征上使用统一管道?

时间:2019-02-12 09:30:39

标签: python machine-learning scikit-learn

要在分类特征上运行编码器,在数字特征上运行Imputer(参见下文),并将它们统一在一起。
例如,具有分类特征的数值:

df_with_cat = pd.DataFrame({
           'A'      : ['ios', 'android', 'web', 'NaN'],
           'B'      : [4, 4, 'NaN', 2], 
           'target' : [1, 1, 0, 0] 
       })
df_with_cat.head()

    A        B  target
----------------------
0   ios      4    1
1   android  4    1
2   web     NaN   0
3   NaN      2    0

我们希望在数值功能上运行Imputer,即用“ most_frequent” /“ median” /“ mean” ==> Pipeline 1 替换缺失值/ NaN。但是我们要将类别特征转换为数字/ OneHotEncoding等==> 管道2

统一它们的最佳实践是什么?
p.s:使用分类器(随机森林/决策树/ GBM)统一上述两种方法

2 个答案:

答案 0 :(得分:1)

正如@Sergey Bushmanov所述,可以使用ColumnTransformer来实现相同的功能。

from sklearn.compose import ColumnTransformer
from sklearn.impute import SimpleImputer
from sklearn.preprocessing import OneHotEncoder

df = pd.DataFrame({
           'A'      : ['ios', 'android', 'web', 'NaN'],
           'B'      : [4, 4, 'NaN', 2], 
           'target' : [1, 1, 0, 0] 
       })

categorical_features = ['A']
numeric_features = ['B']
TARGET = ['target']

df[numeric_features]=df[numeric_features].replace('NaN', np.NaN)
columnTransformer = ColumnTransformer(
    transformers=[
        ('cat', OneHotEncoder(), categorical_features),
        ('num', SimpleImputer( strategy='most_frequent'), numeric_features)])

columnTransformer.fit_transform(df)

#
array([[0., 0., 1., 0., 4.],
   [0., 1., 0., 0., 4.],
   [0., 0., 0., 1., 4.],
   [1., 0., 0., 0., 2.]])

答案 1 :(得分:0)

对于此df,显然有一种很酷的方法!

df_with_cat = pd.DataFrame({
           'A'      : ['ios', 'android', 'web', 'NaN'],
           'B'      : [4, 4, 'NaN', 2], 
           'target' : [1, 1, 0, 0] 
       })

如果您不介意将sklearn升级到0.20.2,请运行:

pip3 install scikit-learn==0.20.2

并使用此解决方案(由@AI_learning建议):

from sklearn.compose import ColumnTransformer
from sklearn.preprocessing import OneHotEncoder

columnTransformer = ColumnTransformer(
    transformers=[
        ('cat', OneHotEncoder(), CATEGORICAL_FEATURES),
        ('num', Imputer( strategy='most_frequent'), NUMERICAL_FEATURES)
    ])

然后:

columnTransformer.fit(df_with_cat)

但是,如果您使用的是较早的sklearn版本,请使用以下版本:

from sklearn.pipeline import Pipeline
from sklearn.preprocessing import Imputer
from sklearn.preprocessing import LabelBinarizer, LabelEncoder 

CATEGORICAL_FEATURES = ['A']
NUMERICAL_FEATURES = ['B']
TARGET = ['target']

numerical_pipline = Pipeline([
    ('selector', DataFrameSelector(NUMERICAL_FEATURES)),
    ('imputer', Imputer(strategy='most_frequent'))
])

categorical_pipeline = Pipeline([
    ('selector', DataFrameSelector(CATEGORICAL_FEATURES)),
    ('cat_encoder', LabelBinarizerPipelineFriendly())
])

如果您引起注意,我们会错过DataFrameSelector,它不是sklearn的一部分,所以我们在这里写下:

from sklearn.base import BaseEstimator, TransformerMixin
class DataFrameSelector(BaseEstimator, TransformerMixin):
    def __init__(self, attribute_names):
        self.attribute_names = attribute_names
    def fit(self, X, y=None):
        return self
    def transform(self, X):
        return X[self.attribute_names].values

让我们统一它们:

from sklearn.pipeline import FeatureUnion, make_pipeline

preprocessing_pipeline = FeatureUnion(transformer_list=[
    ('numerical_pipline', numerical_pipline),
    ('categorical_pipeline', categorical_pipeline)
])

就这样,现在开始运行:

preprocessing_pipeline.fit_transform(df_with_cat[CATEGORICAL_FEATURES+NUMERICAL_FEATURES])

现在让我们变得更加疯狂! 通过分类程序管道统一它们:

from sklearn import tree
clf = tree.DecisionTreeClassifier()
full_pipeline = make_pipeline(preprocessing_pipeline, clf)

并一起训练他们:

full_pipeline.fit(df_with_cat[CATEGORICAL_FEATURES+NUMERICAL_FEATURES], df_with_cat[TARGET])

只需打开Jupyter笔记本,获取代码片段,然后自己尝试一下!

这是LabelBinarizerPipelineFriendly()的定义:

class LabelBinarizerPipelineFriendly(LabelBinarizer):
    '''
     Wrapper to LabelBinarizer to allow usage in sklearn.pipeline
    '''

    def fit(self, X, y=None):
        """this would allow us to fit the model based on the X input."""
        super(LabelBinarizerPipelineFriendly, self).fit(X)
    def transform(self, X, y=None):
        return super(LabelBinarizerPipelineFriendly, self).transform(X)

    def fit_transform(self, X, y=None):
        return super(LabelBinarizerPipelineFriendly, self).fit(X).transform(X)

这种方法的主要优点是,您可以将经过训练的模型和所有管道都转储到pkl文件中,然后可以实时使用完全相同的模型(在生产中进行预测)