OneHotEncoder ValueError：输入包含 NaN

Question

我已经下载了this 数据，这是我的代码：

from sklearn.metrics import classification_report, confusion_matrix
from sklearn.utils.multiclass import unique_labels
import plotly.figure_factory as ff
import pandas as pd
from sklearn.preprocessing import LabelEncoder, StandardScaler
from sklearn.impute import SimpleImputer
import numpy as np
from sklearn.impute import KNNImputer
from sklearn.model_selection import train_test_split
from sklearn.svm import SVC
from sklearn.tree import DecisionTreeClassifier
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import OrdinalEncoder, OneHotEncoder
from sklearn.compose import make_column_transformer
random_state = 27912

df_train = pd.read_csv("...")
df_test  = pd.read_csv("...")


X_train, X_test, y_train, y_test = train_test_split(df_train.drop(["Survived", "Ticket", "Cabin", "Name", "PassengerId"], 
                                                                  axis = 1), 
                                                    df_train["Survived"], test_size=0.2, 
                                                    random_state=42)

numeric_col_names = ["Age", "SibSp", "Parch", "Fare"]
ordinal_col_names = ["Pclass"]
one_hot_col_names = ["Embarked", "Sex"]

ct = make_column_transformer(
    (SimpleImputer(strategy="median"), numeric_col_names),
    (SimpleImputer(strategy="most_frequent"), ordinal_col_names + one_hot_col_names),
    (OrdinalEncoder(), ordinal_col_names),
    (OneHotEncoder(), one_hot_col_names),
    (StandardScaler(), ordinal_col_names + one_hot_col_names + numeric_col_names))

preprocessing_pipeline = Pipeline([("transformers", ct)])

preprocessing_pipeline.fit_transform(X_train)

我正在尝试将column_transformer 用于预处理步骤，但是，OneHotEncoding 步骤给了我一个错误，ValueError: Input contains NaN。我真的不知道为什么会发生这种情况，因为我之前在估算值。关于为什么会发生这种情况的任何线索？

尝试这样的事情也无济于事

preprocessing_pipeline = Pipeline([("transformers", ct_first)])
ct_second = make_column_transformer((OneHotEncoder(), one_hot_col_names),(StandardScaler(), ordinal_col_names + one_hot_col_names + numeric_col_names))
pipeline = Pipeline([("transformer1", preprocessing_pipeline), ("transformer2", ct_second)])
pipeline.fit_transform(X_train)

我想知道为什么会发生这种情况以及为什么上面的代码，第一次和第二次尝试都不正确。 谢谢

Answer 1

您需要为每种列类型创建一个管道，以确保按顺序应用不同的步骤（即确保在编码和缩放之前估算缺失值），另请参见 scikit- 中的 this example学习文档。

import pandas as pd
from sklearn.preprocessing import StandardScaler, OrdinalEncoder, OneHotEncoder
from sklearn.impute import SimpleImputer
from sklearn.pipeline import Pipeline
from sklearn.compose import make_column_transformer

# Load the data (from https://www.kaggle.com/c/titanic/data)
df_train = pd.read_csv('train.csv')
df_test = pd.read_csv('test.csv')

# Extract the features
X_train = df_train.drop(labels=['Survived', 'Ticket', 'Cabin', 'Name', 'PassengerId'], axis=1)
X_test = df_test.drop(labels=['Ticket', 'Cabin', 'Name', 'PassengerId'], axis=1)

# Map the feature names to the corresponding 
# types (numerical, ordinal or categorical)
numeric_col_names = ['Age', 'SibSp', 'Parch', 'Fare']
ordinal_col_names = ['Pclass']
one_hot_col_names = ['Embarked', 'Sex']

# Define the numerical features pipeline
numeric_col_transformer = Pipeline([
    ('imputer', SimpleImputer(strategy='median')),
    ('scaler', StandardScaler())
])

# Define the ordinal features pipeline
ordinal_col_transformer = Pipeline([
    ('imputer', SimpleImputer(strategy='most_frequent')),
    ('encoder', OrdinalEncoder()),
    ('scaler', StandardScaler())
])

# Define the categorical features pipeline
one_hot_col_transformer = Pipeline([
    ('imputer', SimpleImputer(strategy='most_frequent')),
    ('encoder', OneHotEncoder(sparse=False)),
    ('scaler', StandardScaler())
])

# Create the overall preprocessing pipeline
preprocessing_pipeline = make_column_transformer(
    (numeric_col_transformer, numeric_col_names),
    (ordinal_col_transformer, ordinal_col_names),
    (one_hot_col_transformer, one_hot_col_names),
)

# Fit the pipeline to the training data
preprocessing_pipeline.fit(X_train)

# Apply the pipeline to the training and test data
X_train_ = preprocessing_pipeline.transform(X_train)
X_test_ = preprocessing_pipeline.transform(X_test)