python - SMOTE 函数在 make_pipeline 中不起作用

Question

我想同时应用交叉验证和过采样。
我从这段代码中得到以下错误:

from sklearn.pipeline import Pipeline, make_pipeline
imba_pipeline = make_pipeline(SMOTE(random_state=42), 
                              LogisticRegression(C=3.4))
cross_val_score(imba_pipeline, X_train_tf, y_train, scoring='f1-weighted', cv=kf)

ll intermediate steps should be transformers and implement fit and transform or be the string 'passthrough' 'SMOTE(k_neighbors=5, kind='deprecated', m_neighbors='deprecated', n_jobs=1, out_step='deprecated', random_state=42, ratio=None, sampling_strategy='auto', svm_estimator='deprecated')' (type ) doesn't

附注。我使用 imblearn.over_sampling.RandomOverSampler 而不是 SMOTE 得到同样的错误。

Answer 1

您应该从 make_pipeline 而不是从 imblearn.pipeline 导入 o​​jit_code :来自 sklearn 的 sklearn.pipeline 需要转换器来实现 make_pipeline 和 fit 方法，但 transform 没有实现 SMOTE 。