himalaya.lasso.solve_sparse_group_lasso_cv

himalaya.lasso.solve_sparse_group_lasso_cv(X, Y, groups=None, l21_regs=[0.05], l1_regs=[0.05], cv=5, max_iter=300, tol=0.0001, momentum=True, n_targets_batch=None, progress_bar=True)

Solves the sparse group Lasso, selecting hyperparameters over cross-validation.

Parameters

Xarray of shape (n_samples, n_features)

Input features.

Yarray of shape (n_samples, n_targets)

Target data.

groupsarray of shape (n_features, ) or None

Encoding of the group of each feature. If None, all features are gathered in one group, and the problem is equivalent to the Lasso.

l21_regsarray of shape (n_l21_regs, )

All the Group Lasso regularization parameter tested.

l1_regsarray of shape (n_l1_regs, )

All the Lasso regularization parameter tested.

cvint or scikit-learn splitter

Cross-validation splitter. If an int, KFold is used.

max_iterint > 0

Maximum number of iterations.

tolfloat > 0

Tolerance of the stopping criterion.

momentumbool

If True, use acceleration in the projected gradient descent.

n_targets_batchint or None

Size of the batch for computing predictions. Used for memory reasons. If None, uses all n_targets at once.

progress_barbool

If True, display a progress bar over batches and iterations.

Returns

coefarray of shape (n_features, n_targets)

Coefficient of the linear model. Always on CPU.

best_l21_regarray of shape (n_targets, )

Best hyperparameter per target.

best_l1_regarray of shape (n_targets, )

Best hyperparameter per target.

all_cv_scoresarray of shape (n_l21_regs * n_l1_regs, n_targets)

Cross-validation scores of all tested hyperparameters.