DocumentationNeurondB Documentation

Hyperparameter Tuning

Bayesian Optimization

Uses probabilistic models to guide search toward promising parameter regions.

Bayesian optimization

SELECT neurondb.bayesian_optimization(
    'random_forest',
    'training_data',
    'features',
    'label',
    jsonb_build_object(
        'n_trees', jsonb_build_object('low', 10, 'high', 100),
        'max_depth', jsonb_build_object('low', 5, 'high', 20)
    ),
    20  -- Number of iterations
) as best_params;

Next Steps