Scikit-Learn GridSearchCV LabelSpreading

Hyperparameter tuning is a crucial step in optimizing machine learning models for best performance. In this example, we’ll demonstrate how to use scikit-learn’s GridSearchCV to perform hyperparameter tuning for LabelSpreading, a semi-supervised learning algorithm.

Grid search is a method for evaluating different combinations of model hyperparameters to find the best performing configuration. It exhaustively searches through a specified parameter grid, trains and evaluates the model for each combination using cross-validation, and selects the hyperparameters that yield the best performance metric.

LabelSpreading is a graph-based semi-supervised learning algorithm that propagates labels through the dataset. It is particularly useful for datasets with a large amount of unlabeled data and a small amount of labeled data.

The key hyperparameters for LabelSpreading include the kernel, which can be either ‘knn’ or ‘rbf’, the gamma parameter for the ‘rbf’ kernel, and n_neighbors for the ‘knn’ kernel.

from sklearn.datasets import make_classification
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.semi_supervised import LabelSpreading
import numpy as np

# Generate synthetic dataset
X, y = make_classification(n_samples=1000, n_features=20, n_informative=15, n_redundant=5, random_state=42)
y[np.random.choice(1000, 800, replace=False)] = -1  # Create a partially labeled dataset

# Split into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Define parameter grid
param_grid = {
    'kernel': ['knn', 'rbf'],
    'gamma': [0.1, 1, 10],  # for 'rbf' kernel
    'n_neighbors': [5, 10, 20]  # for 'knn' kernel
}

# Perform grid search
grid_search = GridSearchCV(estimator=LabelSpreading(),
                           param_grid=param_grid,
                           cv=5,
                           scoring='accuracy')
grid_search.fit(X_train, y_train)

# Report best score and parameters
print(f"Best score: {grid_search.best_score_:.3f}")
print(f"Best parameters: {grid_search.best_params_}")

# Evaluate on test set
best_model = grid_search.best_estimator_
accuracy = best_model.score(X_test, y_test)
print(f"Test set accuracy: {accuracy:.3f}")

Running the example gives an output like:

Best score: 0.176
Best parameters: {'gamma': 0.1, 'kernel': 'rbf', 'n_neighbors': 5}
Test set accuracy: 0.175

The steps are as follows:

1. Generate a synthetic dataset with make_classification, adding a large portion of unlabeled data points.
2. Split the dataset into train and test sets using train_test_split.
3. Define the parameter grid with values for kernel, gamma, and n_neighbors.
4. Perform grid search with GridSearchCV, specifying LabelSpreading, the parameter grid, 5-fold cross-validation, and accuracy as the scoring metric.
5. Report the best cross-validation score and the best set of hyperparameters.
6. Evaluate the best model on the test set and report the accuracy.

By using GridSearchCV, we can efficiently explore different hyperparameter settings and find the combination that maximizes the model’s performance. This automated approach saves time and effort compared to manual hyperparameter tuning and helps ensure we select the best configuration for our LabelSpreading model.