我正在尝试使用scikit-learn中的新管道可视化功能。在jupyter书籍或Google Colab中，我得到的输出是文本而不是管道可视化图表。我希望得到的图表是Scikit-Learn文档中展示的那种。

请提供建议

import seaborn as snsfrom sklearn.pipeline import Pipelinefrom sklearn.impute import SimpleImputerfrom sklearn.preprocessing import StandardScaler, OneHotEncoderfrom sklearn.compose import ColumnTransformerfrom sklearn.linear_model import LogisticRegressionTitanicDataset=sns.load_dataset("titanic")##print(tips.info())##print(tips.head())X=TitanicDataset[["sex","age","fare","embarked","who","pclass"]].copy()y=TitanicDataset[["survived"]].copy()print(X.info())numeric_independent_variables= ['fare', 'age', 'pclass']categorical_independent_variables=["who","embarked","sex"]numeric_pipeline=Pipeline([('imputer', SimpleImputer(strategy='median')),('scaler', StandardScaler())])categorical_pipeline= Pipeline([('imputer', SimpleImputer(strategy='constant', fill_value='missing')),                                ('onehot', OneHotEncoder(handle_unknown='ignore'))])consolidated_pipeline=ColumnTransformer([('num', numeric_pipeline, numeric_independent_variables),        ('cat', categorical_pipeline, categorical_independent_variables)])clf = Pipeline(steps=[('consolidated_pipeline', consolidated_pipeline),                      ('classifier', LogisticRegression())])from sklearn import set_configset_config(display='diagram')print(clf)```

输出

Pipeline(steps=[('consolidated_pipeline',                     ColumnTransformer(transformers=[('num',                                                      Pipeline(steps=[('imputer',                                                                       SimpleImputer(strategy='median')),                                                                      ('scaler',                                                                       StandardScaler())]),                                                      ['fare', 'age', 'pclass']),                                                     ('cat',                                                      Pipeline(steps=[('imputer',                                                                       SimpleImputer(fill_value='missing',                                                                                     strategy='constant')),                                                                      ('onehot',                                                                       OneHotEncoder(handle_unknown='ignore'))]),                                                      ['who', 'embarked',                                                       'sex'])])),                    ('classifier', LogisticRegression())])

回答：

不要使用print。print总是使用对象的__repr__。相反，请使用IPython.display中的display，或者在笔记本的单元格末尾直接使用clf（而不是print(clf)）。