Python package¶
The t-boost Python package provides the following classes, functions and modules. See Quick start for examples.
Classes¶
TBoostRegressor¶
Class purpose
Training and applying regression models: squared_error, poisson, gamma and tweedie.
TBoostClassifier¶
Class purpose
Training and applying classification models: logistic for two classes, softmax for three or more.
ContributionMatrix¶
Class purpose
The dense result of predict_contributions(return_format="matrix").
Functions¶
recommended_recipe¶
Return an estimator configured with the benchmark recipe.
Modules¶
metrics¶
Deviances and weight-aware ranking metrics, calculated separately from the training.
Exceptions and warnings¶
TBoostError and its subclasses, PrecisionWarning¶
The errors raised by t-boost, and the warning issued when features are converted to float32.