mojolearn.AdditiveChi2Sampler#

class mojolearn.AdditiveChi2Sampler(*, sample_steps=2, sample_interval=None)[source]#

Bases: _XNeighbors

Approximate feature map for the additive chi-squared kernel.

A deterministic Fourier sampling of the kernel’s feature map, following scikit-learn’s AdditiveChi2Sampler. Each input feature becomes 2 * sample_steps - 1 output features. Input must be non-negative; negative values raise ValueError. A scipy.sparse input gives a sparse (CSR) output. Computation is float32.

Parameters:

sample_steps (int, default 2) – Number of sampling points.
sample_interval (float or None, default None) – Sampling interval. Required when sample_steps is not 1, 2 or 3 (ValueError otherwise); by default 0.8, 0.5 or 0.4 for those.
numeric_mode ({‘identical’, ‘deterministic’, ‘fast’} or None, default None) – Numeric mode for this estimator ('fast' is Apple only). None uses the process default.

Variables:

n_features_in_ (int) – Number of features seen in fit.

Examples

>>> import numpy as np, mojolearn
>>> X = np.random.default_rng(0).random((1000, 8), dtype=np.float32)
>>> Z = mojolearn.AdditiveChi2Sampler(sample_steps=2).fit_transform(X)

__init__(*, sample_steps=2, sample_interval=None)[source]#

Methods

`fit`(X[, y])	Validate the input and the sampling settings.
`fit_transform`(X[, y])	Fit to X, then transform X: `fit(X).transform(X)`.
`get_feature_names_out`([input_features])	Output feature names.
`get_params`([deep])
`numeric_mode_used`()	The numeric mode this estimator runs in.
`set_params`(**params)	Set constructor parameters.
`transform`(X)	Apply the approximate feature map to X.
`vendor_used`()	The GPU backend this estimator runs on.

fit(X, y=None)[source]#

Validate the input and the sampling settings.

Parameters:

X (array-like or scipy.sparse matrix of shape (n_samples, n_features)) – Training data; must be non-negative (ValueError otherwise).
y (None) – Ignored.

Returns:

self (AdditiveChi2Sampler) – The fitted transformer.

transform(X)[source]#

Apply the approximate feature map to X.

Parameters:: X (array-like or scipy.sparse matrix of shape (n_samples, n_features)) – Non-negative samples (ValueError otherwise).
Returns:: X_new (Array or scipy.sparse.csr_matrix of shape (n_samples, n_features * (2 * sample_steps - 1))) – float32; sparse (CSR) when X is sparse.

fit_transform(X, y=None)[source]#

Fit to X, then transform X: fit(X).transform(X).

Parameters:

X (array-like or scipy.sparse matrix of shape (n_samples, n_features)) – Non-negative training data.
y (None) – Ignored.

Returns:

X_new (Array or scipy.sparse.csr_matrix of shape (n_samples, n_features * (2 * sample_steps - 1)))

get_feature_names_out(input_features=None)[source]#

Output feature names.

For each input name f: additivechi2sampler_f_sqrt, then for each step j from 1 the ..._cos{j} and ..._sin{j} names.

Parameters:: input_features (list of str or None, default None) – Input feature names; None means x0, x1, … Its length must equal n_features_in_ (ValueError otherwise).
Returns:: names (list of str)