skglm.solvers.GroupBCD

class skglm.solvers.GroupBCD(max_iter=1000, max_epochs=100, p0=10, tol=0.0001, fit_intercept=False, warm_start=False, ws_strategy='subdiff', verbose=0)

Block coordinate descent solver for group problems.

Attributes:

w_initarray, shape (n_features,), default None

Initial value of coefficients. If set to None, a zero vector is used instead.

Xw_initarray, shape (n_samples,), default None

Initial value of model fit. If set to None, a zero vector is used instead.

p0int, default 10

Minimum number of groups to be included in the working set.

max_iterint, default 1000

Maximum number of iterations.

max_epochsint, default 100

Maximum number of epochs.

tolfloat, default 1e-4

Tolerance for convergence.

verbosebool, default False

Amount of verbosity. 0/False is silent.

__init__(max_iter=1000, max_epochs=100, p0=10, tol=0.0001, fit_intercept=False, warm_start=False, ws_strategy='subdiff', verbose=0)

Methods

__init__([max_iter, max_epochs, p0, tol, ...])
custom_checks(X, y, datafit, penalty)	Ensure the solver is suited for the datafit + penalty problem.
solve(X, y, datafit, penalty[, w_init, ...])	Solve the optimization problem after validating its compatibility.