New error message from established workflow: AttributeError: 'Pipeline' object has no attribute '_check_fit_params'[BUG]

[dysartk]

Describe the bug

I received this error message when attempting to run an established workflow that is dependent on creating a pipeline, applying SMOTE, and then using GridSearchCV to tune the hyperparameters. I've replicated the error when running the code on both a machine running the latest version of PopOS and Mac OS X. Python version 3.11.5. I've pasted a sample code below, generated with Copilot, that replicates the exact error and pipeline build. This code returned no errors in the last 24 hours from this post.

Steps/Code to Reproduce

from imblearn.over_sampling import SMOTE
from imblearn.pipeline import Pipeline
from sklearn import svm
from sklearn.datasets import make_classification
from sklearn.model_selection import GridSearchCV, train_test_split
from sklearn.metrics import classification_report

# Create a binary classification dataset
X, y = make_classification(n_samples=1000, n_features=20, n_informative=2, n_redundant=10, n_clusters_per_class=1, weights=[0.99], flip_y=0, random_state=1)

# Split the dataset into training set and test set
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Define pipeline
model = svm.SVC()
resampling = SMOTE(sampling_strategy=0.5) # SMOTE happens during Cross Validation not before..
pipeline = Pipeline([('SMOTE', resampling), ('SVM', model)])

# Define parameter grid
param_grid = [
    {'SVM__C': [1, 10, 100, 1000], 'SVM__kernel': ['linear']},
    {'SVM__C': [1, 10, 100, 1000], 'SVM__gamma': [0.001, 0.0001], 'SVM__kernel': ['rbf']},
]

# Apply Grid Search
grid_search = GridSearchCV(pipeline, param_grid, cv=5, scoring='accuracy')
grid_search.fit(X_train, y_train)

# Print the best parameters
print("Best parameters: ", grid_search.best_params_)

# Predict on the test set
y_pred = grid_search.predict(X_test)

# Print the classification report
print(classification_report(y_test, y_pred))

Error Message

ValueError                                Traceback (most recent call last)
Cell In[1], [line 27](vscode-notebook-cell:?execution_count=1&line=27)
     [25](vscode-notebook-cell:?execution_count=1&line=25) # Apply Grid Search
     [26](vscode-notebook-cell:?execution_count=1&line=26) grid_search = GridSearchCV(pipeline, param_grid, cv=5, scoring='accuracy')
---> [27](vscode-notebook-cell:?execution_count=1&line=27) grid_search.fit(X_train, y_train)
     [29](vscode-notebook-cell:?execution_count=1&line=29) # Print the best parameters
     [30](vscode-notebook-cell:?execution_count=1&line=30) print("Best parameters: ", grid_search.best_params_)

File [~/anaconda3/lib/python3.11/site-packages/sklearn/base.py:1474](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/base.py:1474), in _fit_context.<locals>.decorator.<locals>.wrapper(estimator, *args, **kwargs)
   [1467](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/base.py:1467)     estimator._validate_params()
   [1469](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/base.py:1469) with config_context(
   [1470](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/base.py:1470)     skip_parameter_validation=(
   [1471](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/base.py:1471)         prefer_skip_nested_validation or global_skip_validation
   [1472](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/base.py:1472)     )
   [1473](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/base.py:1473) ):
-> [1474](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/base.py:1474)     return fit_method(estimator, *args, **kwargs)

File [~/anaconda3/lib/python3.11/site-packages/sklearn/model_selection/_search.py:970](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/model_selection/_search.py:970), in BaseSearchCV.fit(self, X, y, **params)
    [964](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/model_selection/_search.py:964)     results = self._format_results(
    [965](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/model_selection/_search.py:965)         all_candidate_params, n_splits, all_out, all_more_results
    [966](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/model_selection/_search.py:966)     )
    [968](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/model_selection/_search.py:968)     return results
--> [970](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/model_selection/_search.py:970) self._run_search(evaluate_candidates)
    [972](https://file+.vscode-resource.vscode-cdn.net/home/kevindysart/Desktop/~/anaconda3/lib/python3.11/site-packages/sklearn/model_selection/_search.py:972) # multimetric is determined here because in the case of a callable
...
  File "/home/kevindysart/anaconda3/lib/python3.11/site-packages/imblearn/pipeline.py", line 292, in fit
    fit_params_steps = self._check_fit_params(**fit_params)
                       ^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'Pipeline' object has no attribute '_check_fit_params'

[glemaitre]

What is your version of scikit-learn?

[glemaitre]

I assume that a quick fix would be to upgrade scikit-learn. I'll check on my side if we have a compatibility issue with older versions.

[mahanrahimi]

Are there any updates on this issue? I have the same problem of getting AttributeError: 'Pipeline' object has no attribute '_check_fit_params' on Windows with:
python = 3.10.14
scikit-learn = 1.4.2
imbalanced-learn = 0.11.0
joblib = 1.4.0

[dysartk]

Sorry for the delay. Ultimately, the solution for me was similar to your suggestions. I needed to uninstall and re-install Anaconda and select that environment within VS Code after confirming the version of scikit-learn. Thanks, Kevin From: Guillaume Lemaitre ***@***.***> Date: Friday, April 19, 2024 at 1:09 PM To: scikit-learn-contrib/imbalanced-learn ***@***.***> Cc: Dysart, Kevin ***@***.***>, Author ***@***.***> Subject: Re: [scikit-learn-contrib/imbalanced-learn] New error message from established workflow: AttributeError: 'Pipeline' object has no attribute '_check_fit_params'[BUG] (Issue #1078) I assume that a quick fix would be to upgrade scikit-learn. I'll check on my side if we have a compatibility issue with older versions. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because ZjQcmQRYFpfptBannerStart EXTERNAL EMAIL USE CAUTION Do not click links or open attachments unless it is normal to receive them from this sender and you are expecting them. If you do not know who this is from or were not expecting the email please delete the email. Report Suspicious <https://us-phishalarm-ewt.proofpoint.com/EWT/v1/Li743BnW!RyCfkHxZ8eJsBkXB6OaeS-1ZLRpeOvKsTk1EYBqIdfefZ9VTLe9g2oWOuJn7HtSaeALXtd8vV2GFZB9p7sen5sTurRZi-Gm1u6W9d3pmipqx4Sv-FHJC6AcjCwC5k1l4tenKYw$> ZjQcmQRYFpfptBannerEnd I assume that a quick fix would be to upgrade scikit-learn. I'll check on my side if we have a compatibility issue with older versions. — Reply to this email directly, view it on GitHub<https://urldefense.com/v3/__https:/github.com/scikit-learn-contrib/imbalanced-learn/issues/1078*issuecomment-2066969000__;Iw!!Li743BnW!gbI6-NhW6YsP340eUArdVawPkkHhv0QrLS20_2w6NATJBfAWWpCpegMlz5aeVgeLtw7rpP8t0CVn5cx7CmXDdxPUmBy1$>, or unsubscribe<https://urldefense.com/v3/__https:/github.com/notifications/unsubscribe-auth/ACNF3SACHT6COMTCPROLNM3Y6FFV5AVCNFSM6AAAAABGLING22VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDANRWHE3DSMBQGA__;!!Li743BnW!gbI6-NhW6YsP340eUArdVawPkkHhv0QrLS20_2w6NATJBfAWWpCpegMlz5aeVgeLtw7rpP8t0CVn5cx7CmXDd2i8Q9sG$>. You are receiving this because you authored the thread.Message ID: ***@***.***>

[mahanrahimi]

Kevin, could you please let me know which versions of python, scikit-learn, and imbalanced-learn you are using? I have tried different versions. Also, I have uninstalled and re-installed Anaconda. But the same error still shows up.

[glemaitre]

imbalanced-learn = 0.11.0

@mahanrahimi you need to update to 0.12.2 for compatibility with scikit-learn 1.4.2

[mahanrahimi]

imbalanced-learn = 0.11.0

@mahanrahimi you need to update to 0.12.2 for compatibility with scikit-learn 1.4.2

Thank you so much, this resolved my problem.