Classification

NeuronDB supports multiple classification algorithms.

Logistic Regression

Copy
CREATE TEMP TABLE logreg_model AS
SELECT neurondb.train(
    'default',
    'logistic_regression',
    'training_table',
    'label',
    ARRAY['features'],
    '{}'::jsonb
)::integer AS model_id;

Prediction

Copy
SELECT neurondb.predict(
    (SELECT model_id FROM logreg_model),
    features
) AS prediction
FROM test_table;

Other Algorithms

NeuronDB also supports:

• SVM (Support Vector Machine)
• Naive Bayes
• Decision Trees
• Neural Networks

Learn More

For detailed documentation on classification algorithms, evaluation metrics, hyperparameter tuning, and multi-class classification, visit:

Classification Documentation

Related Topics

• Random Forest - Ensemble classification
• Quality Metrics - Classification metrics