Feature generation and construction module.
The generators are designed to augment the feature space with non-linear relationships:
PolynomialFeaturesGenerator: Higher-order terms.
InteractionFeaturesGenerator: Cross-feature products.
MathematicalFeaturesGenerator: Functional mappings (log, exp).
BinningFeaturesGenerator: Continuous to categorical mappings.
Creating interaction features:
>>> from tuiml.features.generation import InteractionFeaturesGenerator
>>> import numpy as np
>>> X = np.array([[1, 2], [3, 4]])
>>> gen = InteractionFeaturesGenerator()
>>> X_new = gen.fit_transform(X)Discretizing continuous features into bins:
>>> from tuiml.features.generation import BinningFeaturesGenerator
>>> binner = BinningFeaturesGenerator(n_bins=5, strategy='uniform')
>>> X_binned = binner.fit_transform(X)