SMOTE function not working in make_pipeline

Question

I wanna simultaneously apply cross-validation and over-sampling. I get the following error from this code:

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

PS. I get the same error using imblearn.over_sampling.RandomOverSampler rather than SMOTE.

Hypsometer

Answer 1

You should import make_pipeline from imblearn.pipeline and not from sklearn.pipeline: make_pipeline from sklearn needs the transformers to implement fit and transform methods but SMOTE does not implement transform.

Drear

Thx, you've literally saved my day, seriously didn't know this and was stuck in a bad manner – Salman