Scikit-Learn GridSearchCV GammaRegressor

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 GammaRegressor, a model used for regression tasks with positive continuous targets.

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.

GammaRegressor models the target variable as a gamma distribution, which is suitable for positive continuous data. The model is useful in scenarios where the target variable cannot be negative, such as predicting the time until an event occurs or modeling insurance claims.

The key hyperparameters for GammaRegressor include the regularization strength (alpha), which controls the amount of regularization applied to the model to prevent overfitting, and the maximum number of iterations (max_iter), which sets the limit on the number of iterations for the optimization solver.

from sklearn.datasets import make_regression
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.linear_model import GammaRegressor

# Generate synthetic regression dataset with positive targets
X, y = make_regression(n_samples=1000, n_features=10, noise=0.1, random_state=42)
y = y - y.min() + 1  # Shift targets to be positive

# 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 = {
    'alpha': [0.01, 0.1, 1],
    'max_iter': [100, 200, 300]
}

# Perform grid search
grid_search = GridSearchCV(estimator=GammaRegressor(),
                           param_grid=param_grid,
                           cv=5,
                           scoring='neg_mean_squared_error')
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_
test_score = best_model.score(X_test, y_test)
print(f"Test set R^2 score: {test_score:.3f}")

Running the example gives an output like:

Best score: 828.560
Best parameters: {'alpha': 0.1, 'max_iter': 100}
Test set R^2 score: 0.932

The steps are as follows:

1. Generate a synthetic regression dataset using make_regression with positive targets by shifting the target values to be positive.
2. Split the dataset into train and test sets using train_test_split.
3. Define the parameter grid with different values for alpha and max_iter hyperparameters.
4. Perform grid search using GridSearchCV, specifying the GammaRegressor model, parameter grid, 5-fold cross-validation, and negative mean squared error scoring metric.
5. Report the best cross-validation score (negated to positive) and best set of hyperparameters found by grid search.
6. Evaluate the best model on the hold-out test set and report the R^2 score.

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 GammaRegressor model.