cca_zoo.stochastic

Mini-batch CCA methods, for data too large to fit a full-batch gradient step in memory.

---

StochasticCCAEY

StochasticCCAEY(
    latent_dimensions: int = 1,
    center: bool = True,
    c: float = 0.0,
    learning_rate: float = 0.01,
    momentum: float = 0.9,
    batch_size: int | None = None,
    max_iter: int = 1000,
    tol: float = 1e-06,
    random_state: int | None = None,
)

Bases: CCAEY

Eckart-Young CCA fit by mini-batch momentum SGD, for large-scale data.

Identical objective to :class:~cca_zoo.linear.gradient.CCAEY (same c ridge blend towards :class:~cca_zoo.linear.gradient.PLSEY, same analytic gradient) but fit the way :class:~sklearn.linear_model.SGDRegressor fits a linear model: each epoch, the data is shuffled once and split into batch_size chunks (:func:sklearn.utils.gen_batches), taking one momentum gradient step per chunk. Use this instead of :class:~cca_zoo.linear.gradient.CCAEY when the full dataset does not fit comfortably in memory or a full-batch gradient evaluation is too slow to repeat every iteration.

Parameters:

Name	Type	Description	Default
latent_dimensions	int	Number of latent dimensions. Default is 1.	1
center	bool	Whether to subtract column means. Default True.	True
c	float	Ridge blend in [0, 1] between CCAEY (0) and PLSEY (1). Default is 0; see :class:~cca_zoo.linear.gradient.CCAEY's docstring for the numerical-stability note on high-dimensional data.	0.0
learning_rate	float	Gradient step size. Default is 1e-2.	0.01
momentum	float	Momentum coefficient in [0, 1). Default is 0.9.	0.9
batch_size	int | None	Mini-batch size. None uses the full dataset (one gradient step per epoch).	None
max_iter	int	Number of epochs (full passes over the shuffled data). Default is 1000.	1000
tol	float	Convergence tolerance on the full-dataset objective's change between consecutive epochs. Default is 1e-6.	1e-06
random_state	int | None	Seed for reproducibility.	None

Examples:

>>> import numpy as np
>>> rng = np.random.default_rng(0)
>>> X1 = rng.standard_normal((5000, 200))
>>> X2 = rng.standard_normal((5000, 150))
>>> model = StochasticCCAEY(latent_dimensions=4, batch_size=128, random_state=0)
>>> model = model.fit([X1, X2])

Source code in cca_zoo/stochastic/_stochastic_cca_ey.py

def __init__(
    self,
    latent_dimensions: int = 1,
    center: bool = True,
    c: float = 0.0,
    learning_rate: float = 1e-2,
    momentum: float = 0.9,
    batch_size: int | None = None,
    max_iter: int = 1000,
    tol: float = 1e-6,
    random_state: int | None = None,
) -> None:
    super().__init__(
        latent_dimensions=latent_dimensions,
        center=center,
        c=c,
        max_iter=max_iter,
        tol=tol,
        random_state=random_state,
    )
    self.learning_rate = learning_rate
    self.momentum = momentum
    self.batch_size = batch_size

fit

fit(
    views: list[ArrayLike], y: None = None
) -> StochasticCCAEY

Fit by mini-batch momentum SGD on the EY loss.

Parameters:

Name	Type	Description	Default
views	list[ArrayLike]	List of 2 or more arrays, each (n_samples, n_features_i).	required
y	None	Ignored.	None

Returns:

Name	Type	Description
self	StochasticCCAEY	Fitted estimator.

Raises:

Type	Description
ValueError	If fewer than 2 views are provided.
ValueError	If views have inconsistent numbers of samples.

Source code in cca_zoo/stochastic/_stochastic_cca_ey.py

def fit(self, views: list[ArrayLike], y: None = None) -> StochasticCCAEY:
    """Fit by mini-batch momentum SGD on the EY loss.

    Args:
        views: List of 2 or more arrays, each (n_samples, n_features_i).
        y: Ignored.

    Returns:
        self: Fitted estimator.

    Raises:
        ValueError: If fewer than 2 views are provided.
        ValueError: If views have inconsistent numbers of samples.
    """
    views_: list[np.ndarray] = self._setup_fit(views)
    rng = np.random.default_rng(self.random_state)
    self.weights_ = self._fit_sgd(views_, rng)
    return self