algorithms

Module: algorithms

Inheritance diagram for regreg.algorithms:

digraph inheritance51d2b39f6a { rankdir=LR; size="8.0, 12.0"; "regreg.algorithms.FISTA" [URL="#regreg.algorithms.FISTA",fontname="Vera Sans, DejaVu Sans, Liberation Sans, Arial, Helvetica, sans",fontsize=10,height=0.25,shape=box,style="setlinewidth(0.5)",target="_top",tooltip="The FISTA generalized gradient algorithm"]; "regreg.algorithms.algorithm" -> "regreg.algorithms.FISTA" [arrowsize=0.5,style="setlinewidth(0.5)"]; "regreg.algorithms.algorithm" [URL="#regreg.algorithms.algorithm",fontname="Vera Sans, DejaVu Sans, Liberation Sans, Arial, Helvetica, sans",fontsize=10,height=0.25,shape=box,style="setlinewidth(0.5)",target="_top"]; }

Classes

FISTA

class regreg.algorithms.FISTA(composite)

Bases: regreg.algorithms.algorithm

The FISTA generalized gradient algorithm

__init__(composite)

Initialize self. See help(type(self)) for accurate signature.

alpha = 1.1
attempt_decrease = False
backtrack(itercount)
debug = False
default_tol = 1e-05
fit(tol=None, min_its=None, max_its=None, FISTA=True, start_step=1.0, restart=inf, coef_stop=False, return_objective_hist=True, monotonicity_restart=True, debug=None, prox_control={})

Use the FISTA (or ISTA) algorithm to fit the problem

Parameters

FISTA : bool

use Nesterov weights? If False, this is just gradient descent

start_step : float

used in backtracking. This is the starting value of self.step

restart : int

Restart Nesterov weights every restart iterations. Default is never (np.inf)

coef_stop : bool

Stop based on coefficient changes instead of objective value

return_objective_hist : bool

Return the sequence of objective values?

monotonicity_restart : bool

If True, Nesterov weights are restarted every time the objective value increases

debug : bool

Resets self.debug, which controls whether convergence information is printed

prox_control : dict

A dictionary of arguments for fit(), used when the composite.proximal_step itself is a FISTA problem

Returns

objective_hist : ndarray

A vector of objective values. Only return if return_objective_hist is True.

max_its = 10000
min_its = 5
property output

Return the ‘interesting’ part of the composite arguments. In the regression case, this is the tuple (beta, r).

perform_backtrack = True
step = None
update_working_coefs(proposed_coefs)

algorithm

class regreg.algorithms.algorithm(composite)

Bases: object

__init__(composite)

Initialize self. See help(type(self)) for accurate signature.

alpha = 1.1
attempt_decrease = False
debug = False
default_tol = 1e-05
fit()

Abstract method.

max_its = 10000
min_its = 5
property output

Return the ‘interesting’ part of the composite arguments. In the regression case, this is the tuple (beta, r).

perform_backtrack = True
step = None