preprocessing.repeating_basis_function

preprocessing.repeating_basis_function

Repeating Basis Function transformer for cyclical features.

Classes

Name	Description
RepeatingBasisFunction	Transformer that encodes cyclical features using repeating radial basis functions.

RepeatingBasisFunction

preprocessing.repeating_basis_function.RepeatingBasisFunction(
    n_periods,
    column,
    input_range,
    remainder='drop',
)

Transformer that encodes cyclical features using repeating radial basis functions.

This transformer places Gaussian basis functions across the specified input range and wraps them around to handle periodicity (e.g., day of year, hour of day). It is a simplified implementation to avoid external dependencies like scikit-lego.

Attributes

Name	Type	Description
n_periods	int	Number of basis functions to place.
column	str	Name of the column in the input DataFrame/Series to transform.
input_range	Tuple[int, int]	The range of the input values (min, max).
remainder	str	Policy for remaining columns (currently only ‘drop’ is supported).

Examples

import pandas as pd
import numpy as np
from spotforecast2_safe.preprocessing.repeating_basis_function import RepeatingBasisFunction

X = pd.DataFrame({"hour": [0, 6, 12, 18, 23]})
rbf = RepeatingBasisFunction(n_periods=4, column="hour", input_range=(0, 23))
features = rbf.fit_transform(X)
print(features.shape)
assert features.shape == (5, 4)

(5, 4)

Methods

Name	Description
fit	Fitted transformer (no-op).
transform	Transform the input data into RBF features.

fit

preprocessing.repeating_basis_function.RepeatingBasisFunction.fit(X, y=None)

Fitted transformer (no-op).

Parameters

Name	Type	Description	Default
X	Any	Input data.	required
y	Any	Ignored.	`None`

Returns

Name	Type	Description
self	RepeatingBasisFunction	The fitted transformer.

Examples

import pandas as pd
from spotforecast2_safe.preprocessing.repeating_basis_function import RepeatingBasisFunction

X = pd.DataFrame({"month": [1, 3, 6, 9, 12]})
rbf = RepeatingBasisFunction(n_periods=6, column="month", input_range=(1, 12))
fitted = rbf.fit(X)
print(type(fitted))
assert fitted is rbf

<class 'spotforecast2_safe.preprocessing.repeating_basis_function.RepeatingBasisFunction'>

transform

preprocessing.repeating_basis_function.RepeatingBasisFunction.transform(X)

Transform the input data into RBF features.

Parameters

Name	Type	Description	Default
X	Union[pd.Series, pd.DataFrame]	Input DataFrame or Series containing the column to transform.	required

Returns

Name	Type	Description
	np.ndarray	np.ndarray: Array of transformed features with shape (n_samples, n_periods).

Raises

Name	Type	Description
	ValueError	If the specified column is not found in the input.

Examples

import pandas as pd
import numpy as np
from spotforecast2_safe.preprocessing.repeating_basis_function import RepeatingBasisFunction

X = pd.DataFrame({"hour": list(range(0, 24, 4))})
rbf = RepeatingBasisFunction(n_periods=4, column="hour", input_range=(0, 23))
rbf.fit(X)
features = rbf.transform(X)
print(f"shape: {features.shape}, dtype: {features.dtype}")
assert features.shape == (6, 4)
assert features.dtype == np.float64
assert np.all((features >= 0) & (features <= 1))

shape: (6, 4), dtype: float64