Scikit-Learn RandomizedSearchCV GammaRegressor

Hyperparameter tuning is essential for optimizing machine learning models. In this example, we’ll demonstrate how to use scikit-learn’s RandomizedSearchCV for hyperparameter tuning of a GammaRegressor model, useful for regression tasks with positive continuous targets.

Random search is a method for evaluating different combinations of model hyperparameters. Unlike grid search, it samples a fixed number of hyperparameter combinations from a specified distribution, making it more efficient when searching over a large hyperparameter space.

Gamma Regressor is designed for regression tasks where the target variable follows a Gamma distribution, commonly used in insurance and other fields dealing with skewed data. The model is trained by minimizing the Gamma loss function.

Key hyperparameters for GammaRegressor include the regularization strength (alpha), which controls model complexity and helps prevent overfitting; and the solver algorithm (solver), which is the optimization method used to find the model coefficients.

from sklearn.datasets import make_regression
from sklearn.model_selection import train_test_split, RandomizedSearchCV
from sklearn.experimental import enable_hist_gradient_boosting
from sklearn.linear_model import GammaRegressor
from scipy.stats import uniform

# Generate synthetic regression dataset
X, y = make_regression(n_samples=1000, n_features=10, noise=0.1, random_state=42)
y = y - y.min() + 1  # Shift target 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 hyperparameter distribution
param_dist = {
    'alpha': uniform(loc=0, scale=10),
    'solver': ['auto', 'lbfgs']
}

# Perform random search
random_search = RandomizedSearchCV(estimator=GammaRegressor(),
                                   param_distributions=param_dist,
                                   n_iter=100,
                                   cv=5,
                                   scoring='neg_mean_gamma_deviance',
                                   random_state=42)
random_search.fit(X_train, y_train)

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

# Evaluate on test set
best_model = random_search.best_estimator_
accuracy = best_model.score(X_test, y_test)
print(f"Test set R^2 score: {accuracy:.3f}")

Running the example gives an output like:

Best score: -0.018
Best parameters: {'alpha': 0.007787658410143283, 'solver': 'lbfgs'}
Test set R^2 score: 0.934

The steps are as follows:

1. Generate a synthetic regression dataset using scikit-learn’s make_regression function and ensure the target is positive.
2. Split the dataset into train and test sets using train_test_split.
3. Define a hyperparameter distribution with different values for alpha and solver hyperparameters.
4. Perform random search using RandomizedSearchCV, specifying the GammaRegressor model, hyperparameter distribution, 100 iterations, 5-fold cross-validation, and negative mean Gamma deviance scoring metric.
5. Report the best cross-validation score and best set of hyperparameters found by random search.
6. Evaluate the best model on the hold-out test set and report the R^2 score.

By using RandomizedSearchCV, 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.