pyccea.projection package

Submodules

pyccea.projection.cipls module

class pyccea.projection.cipls.CIPLS(n_components=10, copy=True)

Bases: BaseEstimator

Covariance-free Partial Least Squares (CIPLS).

Jordao, Artur, et al. “Covariance-free partial least squares: An incremental dimensionality reduction method.” Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (2021). Source: https://github.com/arturjordao/IncrementalDimensionalityReduction

Attributes:

n: int

Number of iterations. It starts at 0 and incrementally goes up to the number of samples (n_samples).

n_features: int

Number of variables.

x_weights_: np.ndarray (n_features, n_components)

Projection matrix.

x_scores_: np.ndarray (n_samples, n_components)

The transformed training samples (latent components).

x_loadings_: np.ndarray (n_features, n_components)

The loadings of X.

y_loadings_: np.ndarray (n_targets, n_components)

The loadings of Y, where n_targets is the number of response variables.

x_rotations_: np.ndarray (n_components, n_features)

Transposed and non-normalized projection matrix.

sum_x: np.ndarray (n_features,)

The sum of each feature individually across all training samples.

sum_y: np.ndarray (1,)

The sum of targets across all training samples.

Methods

fit(X, Y)	Fit model to data
get_metadata_routing()	Get metadata routing of this object.
get_params([deep])	Get parameters for this estimator.
normalize(x)	Scale input vectors individually to unit norm (vector length).
set_params(**params)	Set the parameters of this estimator.
transform(X[, Y])	Apply the dimension reduction learned on the training data.

fit(X, Y)

Fit model to data

Parameters:

X: np.ndarray (n_samples, n_features)

Training data.

Y: np.ndarray (n_samples,) or (n_samples, n_targets)

Target data.

normalize(x)

Scale input vectors individually to unit norm (vector length).

transform(X, Y=None)

Apply the dimension reduction learned on the training data.

Parameters:

X: np.ndarray (n_samples, n_features)

Training data.

Y: np.ndarray (n_samples,) or (n_samples, n_targets), default None

Target data.

pyccea.projection.kpls module

class pyccea.projection.kpls.KernelPLS(n_components: int = 2, kernel: str = 'rbf', gamma: float | None = None, degree: int = 3, coef0: float = 1.0)

Bases: BaseEstimator

Kernel Partial Least Squares (KPLS).

This class implements a kernelized version of Partial Least Squares (PLS), allowing the use of nonlinear mappings through kernel functions such as RBF, polynomial, or linear. The algorithm extracts latent components in the kernel space that maximize covariance between the predictor variables and response(s).

Methods

fit(X, Y)	Fit the KernelPLS model to the training data.
get_metadata_routing()	Get metadata routing of this object.
get_params([deep])	Get parameters for this estimator.
predict(X_new)	Predict using the KernelPLS model.
set_params(**params)	Set the parameters of this estimator.
set_predict_request(*[, X_new])	Request metadata passed to the predict method.

fit(X: ndarray, Y: ndarray)

Fit the KernelPLS model to the training data.

Parameters:

Xndarray of shape (n_samples, n_features)

Training input data.

Yndarray of shape (n_samples,) or (n_samples, n_targets)

Training target data.

Returns:

selfobject

Fitted model.

predict(X_new: ndarray) → ndarray

Predict using the KernelPLS model.

Parameters:

X_newndarray of shape (n_samples_new, n_features)

Input samples.

Returns:

Y_predndarray of shape (n_samples_new, n_targets)

Predicted target values.

set_predict_request(*, X_new: bool | None | str = '$UNCHANGED$') → KernelPLS

Request metadata passed to the predict method.

Note that this method is only relevant if enable_metadata_routing=True (see sklearn.set_config()). Please see User Guide on how the routing mechanism works.

The options for each parameter are:

• True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it to predict.

• None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

• str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Note

This method is only relevant if this estimator is used as a sub-estimator of a meta-estimator, e.g. used inside a Pipeline. Otherwise it has no effect.

Parameters:

X_newstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED

Metadata routing for X_new parameter in predict.

Returns:

selfobject

The updated object.

pyccea.projection.vip module

class pyccea.projection.vip.VIP(model)

Bases: object

Variable Importance in Projection (VIP).

Mehmood, Tahir, et al. “A review of variable selection methods in partial least squares regression.” Chemometrics and intelligent laboratory systems 118 (2012): 62-69. Source: https://github.com/scikit-learn/scikit-learn/issues/7050

Attributes:

n_featuresint

Number of variables.

n_componentsint

Number of components.

x_rotations_np.ndarray (n_features, n_components)

Projection matrix used to transform X.

x_scores_np.ndarray (n_samples, n_components)

The transformed training samples (latent components).

y_loadings_np.ndarray (n_targets, n_components)

The loadings of Y.

importancesnp.ndarray (n_features,)

Importance of each feature based on its contribution to yield the latent space.

Methods

compute()

Calculate feature importances.

compute()

Calculate feature importances.