statsmodels.base.optimizer._fit_lbfgs¶

statsmodels.base.optimizer._fit_lbfgs(f, score, start_params, fargs, kwargs, disp=True, maxiter=100, callback=None, retall=False, full_output=True, hess=None)[source]¶

Fit using Limited-memory Broyden-Fletcher-Goldfarb-Shannon algorithm.

Parameters:¶

f : function¶

Returns negative log likelihood given parameters.

score : function¶

Returns gradient of negative log likelihood with respect to params.

start_params : array_like, optional¶

Initial guess of the solution for the loglikelihood maximization. The default is an array of zeros.

fargs : tuple¶

Extra arguments passed to the objective function, i.e. objective(x,*args)

kwargs : dict[str, Any]¶

Extra keyword arguments passed to the objective function, i.e. objective(x,**kwargs)

disp : bool¶

Set to True to print convergence messages.

maxiter : int¶

The maximum number of iterations to perform.

callback : callable callback(xk)¶

Called after each iteration, as callback(xk), where xk is the current parameter vector.

retall : bool¶

Set to True to return list of solutions at each iteration. Available in Results object’s mle_retvals attribute.

full_output : bool¶

Set to True to have all available output in the Results object’s mle_retvals attribute. The output is dependent on the solver. See LikelihoodModelResults notes section for more information.

hess : str, optional¶

Method for computing the Hessian matrix, if applicable.

Returns:¶

• xopt (ndarray) – The solution to the objective function

• retvals (dict, None) – If full_output is True then this is a dictionary which holds information returned from the solver used. If it is False, this is None.

Notes

Within the mle part of statsmodels, the log likelihood function and its gradient with respect to the parameters do not have notationally consistent sign.