Use sklearn's GridSearchCV with a pipeline, preprocessing just once

Question

I'm using scickit-learn to tune a model hyper-parameters. I'm using a pipeline to have chain the preprocessing with the estimator. A simple version of my problem would look like this:

import numpy as np
from sklearn.model_selection import GridSearchCV
from sklearn.pipeline import make_pipeline
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LogisticRegression


grid = GridSearchCV(make_pipeline(StandardScaler(), LogisticRegression()),
                    param_grid={'logisticregression__C': [0.1, 10.]},
                    cv=2,
                    refit=False)

_ = grid.fit(X=np.random.rand(10, 3),
             y=np.random.randint(2, size=(10,)))

In my case the preprocessing (what would be StandardScale() in the toy example) is time consuming, and I'm not tuning any parameter of it.

So, when I execute the example, the StandardScaler is executed 12 times. 2 fit/predict * 2 cv * 3 parameters. But every time StandardScaler is executed for a different value of the parameter C, it returns the same output, so it'd be much more efficient, to compute it once, and then just run the estimator part of the pipeline.

I can manually split the pipeline between the preprocessing (no hyper parameters tuned) and the estimator. But to apply the preprocessing to the data, I should provide the training set only. So, I would have to implement the splits manually, and not use GridSearchCV at all.

Is there a simple/standard way to avoid repeating the preprocessing while using GridSearchCV?

Answer 1

Update: Ideally, the answer below should not be used as it leads to data leakage as discussed in comments. In this answer, GridSearchCV will tune the hyperparameters on the data already preprocessed by StandardScaler, which is not correct. In most conditions that should not matter much, but algorithms which are too sensitive to scaling will give wrong results.

---

Essentially, GridSearchCV is also an estimator, implementing fit() and predict() methods, used by the pipeline.

So instead of:

grid = GridSearchCV(make_pipeline(StandardScaler(), LogisticRegression()),
                    param_grid={'logisticregression__C': [0.1, 10.]},
                    cv=2,
                    refit=False)

Do this:

clf = make_pipeline(StandardScaler(), 
                    GridSearchCV(LogisticRegression(),
                                 param_grid={'logisticregression__C': [0.1, 10.]},
                                 cv=2,
                                 refit=True))

clf.fit()
clf.predict()

What it will do is, call the StandardScalar() only once, for one call to clf.fit() instead of multiple calls as you described.

Edit:

Changed refit to True, when GridSearchCV is used inside a pipeline. As mentioned in documentation:

refit : boolean, default=True Refit the best estimator with the entire dataset. If “False”, it is impossible to make predictions using this GridSearchCV instance after fitting.

If refit=False, clf.fit() will have no effect because the GridSearchCV object inside the pipeline will be reinitialized after fit(). When refit=True, the GridSearchCV will be refitted with the best scoring parameter combination on the whole data that is passed in fit().

So if you want to make the pipeline, just to see the scores of the grid search, only then the refit=False is appropriate. If you want to call the clf.predict() method, refit=True must be used, else Not Fitted error will be thrown.

Answer 2

For those who stumbled upon a little bit different problem, that I had as well.

Suppose you have this pipeline:

classifier = Pipeline([
    ('vectorizer', CountVectorizer(max_features=100000, ngram_range=(1, 3))),
    ('clf', RandomForestClassifier(n_estimators=10, random_state=SEED, n_jobs=-1))])

Then, when specifying parameters you need to include this 'clf_' name that you used for your estimator. So the parameters grid is going to be:

params={'clf__max_features':[0.3, 0.5, 0.7],
        'clf__min_samples_leaf':[1, 2, 3],
        'clf__max_depth':[None]
        }

Answer 3

It is not possible to do this in the current version of scikit-learn (0.18.1). A fix has been proposed on the github project:

https://github.com/scikit-learn/scikit-learn/issues/8830

https://github.com/scikit-learn/scikit-learn/pull/8322

Answer 4

I joined the party late, but I brought a new solution/insight using Pipeline():

• sub-pipeline containing your model (regression/classifier) as a single component
• main pipeline made of routine components:
  • pre-processing component e.g., scaler, dimension reduction, etc.
  • your refitted GridSearchCV(regressor, param) with desired/best params for your model (Note: don't forget to refit=True) based on @Vivek Kumar remark ref

#build an end-to-end pipeline, and supply the data into a regression model and train and fit within the main pipeline.
#It avoids leaking the test\val-set into the train-set
# Create the sub-pipeline

#create and train the sub-pipeline
from sklearn.linear_model import SGDRegressor
from sklearn.compose import TransformedTargetRegressor
from sklearn.pipeline import Pipeline
from sklearn.model_selection import GridSearchCV

sgd_subpipeline = Pipeline(steps=[#('scaler', MinMaxScaler()), # better to not rescale internally
                                  ('SGD',    SGDRegressor(random_state=0)),
])

# Define the hyperparameter grid
param_grid = {
    'SGD__loss':     ['squared_error', 'epsilon_insensitive', 'squared_epsilon_insensitive', 'huber'],
    'SGD__penalty':  ['l2', 'l1', 'elasticnet'],
    'SGD__alpha':    [0.0001, 0.001, 0.01],
    'SGD__l1_ratio': [0.15, 0.25, 0.5]
}

# Perform grid search
grid_search = GridSearchCV(sgd_subpipeline, param_grid, cv=5, n_jobs=-1, verbose=True, refit=True)
grid_search.fit(X_train, y_train)

# Get the best model
best_sgd_reg = grid_search.best_estimator_

# Print the best hyperparameters
print('=========================================[Best Hyperparameters info]=====================================')
print(grid_search.best_params_)

# summarize best
print('Best MAE: %.3f'  % grid_search.best_score_)
print('Best Config: %s' % grid_search.best_params_)
print('==========================================================================================================')

# Create the main pipeline by chaining refitted GridSerachCV sub-pipeline

sgd_pipeline = Pipeline(steps=[('scaler', MinMaxScaler()), # better to rescale externally
                               ('SGD',    grid_search),
])

# Fit the best model on the training data within pipeline (like fit any model/transformer): pipe.fit(traindf[features], traindf[labels]) #X, y

sgd_pipeline.fit(X_train, y_train)

#--------------------------------------------------------------
# Displaying a Pipeline with a Preprocessing Step and Regression
from sklearn import set_config
set_config(display="text")

---

Alternatively, you can use TransformedTargetRegressor (specifically if you need to descale y as @mloning commented here) and chain this component, including your regression model ref. Note:

• you don't need to set transform argument unless you need descaling; please then check to related posts 1, 2, 3, 4, its score
• Pay attention to this remark about not scaling here since:

... With scaling y you actually lose your units....

• Here, It is recommended to:

... Do the transformation outside the pipeline. ...

#build an end-to-end pipeline, and supply the data into a regression model and train and fit within main pipeline.
#It avoids leaking the test\val-set into the train-set
# Create the sub-pipeline
from sklearn.linear_model import SGDRegressor
from sklearn.pipeline import Pipeline
from sklearn.model_selection import GridSearchCV

sgd_subpipeline = Pipeline(steps=[#('scaler', MinMaxScaler()), # better to not rescale internally
                                  ('SGD',    SGDRegressor(random_state=0)),
])

# Define the hyperparameter grid
param_grid = {
    'SGD__loss':     ['squared_error', 'epsilon_insensitive', 'squared_epsilon_insensitive', 'huber'],
    'SGD__penalty':  ['l2', 'l1', 'elasticnet'],
    'SGD__alpha':    [0.0001, 0.001, 0.01],
    'SGD__l1_ratio': [0.15, 0.25, 0.5]
}

# Perform grid search
grid_search = GridSearchCV(sgd_subpipeline, param_grid, cv=5, n_jobs=-1, verbose=True, refit=True)
grid_search.fit(X_train, y_train)

# Get the best model
best_sgd_reg = grid_search.best_estimator_

# Print the best hyperparameters
print('=========================================[Best Hyperparameters info]=====================================')
print(grid_search.best_params_)

# summarize best
print('Best MAE: %.3f'  % grid_search.best_score_)
print('Best Config: %s' % grid_search.best_params_)
print('==========================================================================================================')



# Create the main pipeline using sub-pipeline made of TransformedTargetRegressor component
from sklearn.compose import TransformedTargetRegressor

TTR_sgd_pipeline = Pipeline(steps=[('scaler', MinMaxScaler()), # better to rescale externally
                                   #('SGD', SGDRegressor()),
                                    ('TTR', TransformedTargetRegressor(regressor= grid_search, #SGDRegressor(),
                                                                       #transformer=MinMaxScaler(),
                                                                       #func=np.log,
                                                                       #inverse_func=np.exp,
                                                                       check_inverse=False))
])



# Fit the best model on the training data within pipeline (like fit any model/transformer): pipe.fit(traindf[features], traindf[labels]) #X, y
#best_sgd_pipeline.fit(X_train, y_train)
TTR_sgd_pipeline.fit(X_train, y_train)

#--------------------------------------------------------------
# Displaying a Pipeline with a Preprocessing Step and Regression
from sklearn import set_config
set_config(display="diagram")

---