scikit-learn Cookbook , Second Edition - Second Edition

Perform a grid search

There are three required elements for a grid search:

• An estimator
• A parameter grid
• A cross-validation scheme

7. We have those three elements. Set up the grid search. Run it on the training set:

from sklearn.model_selection import GridSearchCVgrid_cv = GridSearchCV(svm_est, param_grid=param_grid, cv=cv)grid_cv.fit(X_train, y_train)

8. Look up the best parameters found with the grid search:

grid_cv.best_params_{'svc__C': 10, 'svc__gamma': 0.1}

9. Look up the best score, that pertains to the best estimator:

grid_cv.best_score_0.71250000000000002