[ENH] scikit-learn compatible GAMRegressor and GAMClassifier classes

[nickcorona]

Description

This pull request introduces two new classes, GAMRegressor and GAMClassifier, to the sklearn_api.py file. These classes provide scikit-learn compatible models using Generalized Additive Models (GAM) from the pygam library. The integration allows seamless use of GAM within scikit-learn's estimator interface, enabling these models to be used in standard machine learning pipelines.

Changes

• Added GAMRegressor class to support regression tasks using GAM.
• Added GAMClassifier class to support classification tasks using GAM.
• Both classes support various configurations, including distribution, link function, terms, interactions, callbacks, fit intercept, maximum iterations, tolerance, verbosity, and additional GAM parameters.
• Included example usage of GAMRegressor in the __main__ block with synthetic data generation, model fitting, and evaluation.

Example Usage

if __name__ == '__main__':
    # Generate synthetic data
    X, y = make_regression(n_samples=100, n_features=3, noise=0.1)
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)

    # Initialize GAMRegressor with 'auto' terms
    model = GAMRegressor(terms='auto', verbose=True)
    model.fit(X_train, y_train)

    # Inspect the generated terms
    print(model.model_.terms)

    # Predict and evaluate
    y_pred = model.predict(X_test)
    print(f"Test RMSE: {model.rmse(X_test, y_test):.4f}")

[codecov]

Codecov Report

Attention: Patch coverage is 92.12963% with 17 lines in your changes missing coverage. Please review.

Project coverage is 94.81%. Comparing base (a6c14e4) to head (0afb8f0).

Files with missing lines	Patch %	Lines
pygam/sklearn_api.py	82.82%	17 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #364      +/-   ##
==========================================
- Coverage   95.00%   94.81%   -0.20%     
==========================================
  Files          22       24       +2     
  Lines        3202     3411     +209     
==========================================
+ Hits         3042     3234     +192     
- Misses        160      177      +17     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[fkiraly]

This is excellent, thanks!

Are you still working on this? If so, it would be great if you could use parametrize_with_checks to test the sklearn estimator, this runs an entire battery of tests.