python - 如何在 python 中的管道中结合 LabelBinarizer 和 OneHotEncoder 来处理分类变量？

Question

过去几天我在 stackoverflow 上查找了正确的教程和问答，但没有找到正确的指南，主要是因为显示 LabelBinarizer 或 OneHotEncoder 用例的示例没有显示它如何合并到管道中，反之亦然。反之亦然。

我有一个包含 4 个变量的数据集:

num1    num2    cate1    cate2
3       4       Cat      1
9       23      Dog      0
10      5       Dog      1

num1 和 num2 是数值变量，cate1 和 cate2 是分类变量。我知道在拟合 ML 算法之前我需要以某种方式对分类变量进行编码，但我不太确定在多次尝试后如何在管道中做到这一点。

from sklearn.pipeline import Pipeline, FeatureUnion, make_pipeline
from sklearn.preprocessing import StandardScaler, OneHotEncoder, LabelBinarizer

# Class that identifies Column type
class Columns(BaseEstimator, TransformerMixin):
    def __init__(self, names=None):
        self.names = names
    def fit (self, X, y=None, **fit_params):
        return self
    def transform(self, X):
        return X[self.names]

# Separate target from training features
y = df['MED']
X = df.drop('MED', axis=1)

X_selected = X.filter(['num1', 'num2', 'cate1', 'cate2'])

# from the selected X, further choose categorical only
X_selected_cat = X_selected.filter(['cate1', 'cate2']) # hand selected since some cat var has value 0, 1

# Find the numerical columns, exclude categorical columns
X_num_cols = X_selected.columns[X_selected.dtypes.apply(lambda c: np.issubdtype(c, np.number))] # list of numeric column names, automated here
X_cat_cols = X_selected_cat.columns # list of categorical column names, previously hand-slected

# Split data into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X_selected, y, 
                                                    test_size=0.5, 
                                                    random_state=567, 
                                                    stratify=y)

# Pipeline
pipe = Pipeline([
    ("features", FeatureUnion([
        ('numeric', make_pipeline(Columns(names=X_num_cols),StandardScaler())),
        ('categorical', make_pipeline(Columns(names=X_cat_cols)))
    ])),
    ('LR_model', LogisticRegression()),
])

这给了我错误ValueError:无法将字符串转换为 float :'Cat'

用此替换最后第四行

('categorical', make_pipeline(Columns(names=X_cat_cols),OneHotEncoder()))

会给我相同的ValueError:无法将字符串转换为 float :'Cat'。

用此替换最后第四行

('categorical', make_pipeline(Columns(names=X_cat_cols),LabelBinarizer(),OneHotEncoder()))
])),

会给我一个不同的错误TypeError:fit_transform()需要2个位置参数，但给出了3个。

用此替换最后第四行

('numeric', make_pipeline(Columns(names=X_num_cols),LabelBinarizer())),

会给我这个错误TypeError:fit_transform()需要2个位置参数，但给出了3个。

Answer 1

根据 Marcus 的建议，我尝试但无法安装 scikit-learn dev 版本，但发现了类似的东西，名为 category_encoders .

将代码更改为这样即可:

from sklearn.pipeline import Pipeline, FeatureUnion, make_pipeline
from sklearn.preprocessing import StandardScaler, OneHotEncoder, LabelBinarizer
import category_encoders as CateEncoder

# Class that identifies Column type
class Columns(BaseEstimator, TransformerMixin):
    def __init__(self, names=None):
        self.names = names
    def fit (self, X, y=None, **fit_params):
        return self
    def transform(self, X):
        return X[self.names]

# Separate target from training features
y = df['MED']
X = df.drop('MED', axis=1)

X_selected = X.filter(['num1', 'num2', 'cate1', 'cate2'])

# from the selected X, further choose categorical only
X_selected_cat = X_selected.filter(['cate1', 'cate2']) # hand selected since some cat var has value 0, 1

# Find the numerical columns, exclude categorical columns
X_num_cols = X_selected.columns[X_selected.dtypes.apply(lambda c: np.issubdtype(c, np.number))] # list of numeric column names, automated here
X_cat_cols = X_selected_cat.columns # list of categorical column names, previously hand-slected

# Split data into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X_selected, y, 
                                                    test_size=0.5, 
                                                    random_state=567, 
                                                    stratify=y)

# Pipeline
pipe = Pipeline([
    ("features", FeatureUnion([
        ('numeric', make_pipeline(Columns(names=X_num_cols),StandardScaler())),
        ('categorical', make_pipeline(Columns(names=X_cat_cols),CateEncoder.BinaryEncoder()))
    ])),
    ('LR_model', LogisticRegression()),
])