Grid search parameter tuning

1. Introduction

When I read the paper before, I saw grid search for hyperparameter tuning. At first, I thought it should be a relatively tall thing, but in fact, it was a violent search later. If a model has a total of 3 parameters, namely A, B, and C, if A has 3 choices, B has 4 choices, and C has 5 choices, in order to select the best model, you need one parameter and one parameter Try it, from this 60 ( $3\ast 4\ast 5$ ) Find one of the models that meets your needs.

Note: grid search is a way of tuning, and has nothing to do with the way of implementation.

You can either use a simple for loop or directly use the class sklearn.model_selection.GridSearchCV in sklearn .

I think this blog is very good: Necessary for tuning parameters – Grid Search grid search

The shortcomings of grid search are also obvious, and it is time-consuming. Therefore, in practical applications, it is necessary to reduce the number of tuning parameters and narrow the search range of each parameter .

2. Code implementation

The code provided by sklearn's official website is quite good. sklearn official website parameter tuning example

from sklearn import datasets
from sklearn.model_selection import train_test_split
from sklearn.model_selection import GridSearchCV
from sklearn.metrics import classification_report
from sklearn.svm import SVC

print(__doc__)

# Loading the Digits dataset
digits = datasets.load_digits()

# To apply an classifier on this data, we need to flatten the image, to
# turn the data in a (samples, feature) matrix:
n_samples = len(digits.images)
X = digits.images.reshape((n_samples, -1))
y = digits.target

# Split the dataset in two equal parts
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.5, random_state=0)

# Set the parameters by cross-validation
tuned_parameters = [{
    
    'kernel': ['rbf'], 'gamma': [1e-3, 1e-4],
                     'C': [1, 10, 100, 1000]},
                    {
    
    'kernel': ['linear'], 'C': [1, 10, 100, 1000]}]

scores = ['precision', 'recall']

for score in scores:
    print("# Tuning hyper-parameters for %s" % score)
    print()

    clf = GridSearchCV(
        SVC(), tuned_parameters, scoring='%s_macro' % score
    )
    clf.fit(X_train, y_train)

    print("Best parameters set found on development set:")
    print()
    print(clf.best_params_)
    print()
    print("Grid scores on development set:")
    print()
    means = clf.cv_results_['mean_test_score']
    stds = clf.cv_results_['std_test_score']
    for mean, std, params in zip(means, stds, clf.cv_results_['params']):
        print("%0.3f (+/-%0.03f) for %r"
              % (mean, std * 2, params))
    print()

    print("Detailed classification report:")
    print()
    print("The model is trained on the full development set.")
    print("The scores are computed on the full evaluation set.")
    print()
    y_true, y_pred = y_test, clf.predict(X_test)
    print(classification_report(y_true, y_pred))
    print()

# Note the problem is too easy: the hyperparameter plateau is too flat and the
# output model is the same for precision and recall with ties in quality.

3. Others

The official website is in English. If you find it troublesome, you can take a look at this blog: [GridSearchCV, CV tuning hyperparameter usage] [K nearest neighbor classifier KNeighborsClassifier usage] [Cross-validation]