[Machine Learning with Python] Cross Validation and Grid Search: An Example of KNN

Train model:

from sklearn.model_selection import GridSearchCV

param_grid = [
    # try 6 (3×2) combinations of hyperparameters
    {'n_neighbors': [3, 5, 7], 'weights': ['uniform','distance']}
  ]

knn_clf = KNeighborsClassifier()
# train across 3 folds, that's a total of 6*3=18 rounds of training
grid_search = GridSearchCV(knn_clf, param_grid, cv=3,
                           scoring='accuracy', return_train_score=True, n_jobs=-1)
grid_search.fit(X_train, y_train)

Show parameters of best model:

grid_search.best_params_

Show the score of train set:

grid_search.best_score_

Fit on test set:

y_pred = grid_search.predict(X_test)

Show the score of test set:

from sklearn.metrics import accuracy_score
accuracy_score(y_test, y_pred)

More about GridSearchCV: https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.GridSearchCV.html