Scikit-Learn GridSearchCV TweedieRegressor

Tuning hyperparameters is vital for optimizing machine learning models. In this example, we’ll demonstrate how to use scikit-learn’s GridSearchCV to perform hyperparameter tuning for TweedieRegressor, a flexible model for regression tasks.

Grid search is a technique for hyperparameter optimization by exhaustively searching through a parameter grid, training, and evaluating models using cross-validation to find the best configuration.

TweedieRegressor is suitable for a wide range of distributions from normal to Poisson and Gamma. It is particularly useful for datasets with zero-inflated values and continuous distributions.

The most critical hyperparameters for TweedieRegressor include power (the power parameter for the Tweedie distribution), alpha (regularization strength), and max_iter (maximum number of iterations).

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

# Generate synthetic regression dataset
X, y = make_regression(n_samples=1000, n_features=10, noise=0.1, random_state=42)

# 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 = {
    'power': [0, 1, 1.5, 2],
    'alpha': [0.1, 1, 10],
    'max_iter': [100, 1000, 10000]
}

# Perform grid search
grid_search = GridSearchCV(estimator=TweedieRegressor(),
                           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_
score = best_model.score(X_test, y_test)
print(f"Test set R^2 score: {score:.3f}")

Running the example gives an output like:

Best score: -163.719
Best parameters: {'alpha': 0.1, 'max_iter': 100, 'power': 0}
Test set R^2 score: 0.991

The steps are as follows:

1. Generate a synthetic regression dataset using make_regression.
2. Split the dataset into train and test sets using train_test_split.
3. Define the parameter grid with different values for power, alpha, and max_iter hyperparameters.
4. Perform grid search using GridSearchCV, specifying the TweedieRegressor model, parameter grid, 5-fold cross-validation, and negative mean squared error scoring metric.
5. Report the best cross-validation score and best set of hyperparameters found by grid search.
6. Evaluate the best model on the hold-out test set and report the R² score.

By using GridSearchCV, different hyperparameter settings can be explored to find the best-performing configuration for TweedieRegressor. This automated approach streamlines the process of hyperparameter tuning and helps ensure the optimal configuration for regression tasks.