Poisson.fit()

statsmodels.discrete.discrete_model.Poisson.fit

Poisson.fit(start_params=None, method='newton', maxiter=35, full_output=1, disp=1, callback=None, **kwargs) [source]

Fit the model using maximum likelihood.

The rest of the docstring is from statsmodels.base.model.LikelihoodModel.fit

Fit method for likelihood based models

Parameters:

start_params : array-like, optional Initial guess of the solution for the loglikelihood maximization. The default is an array of zeros. method : str, optional The method determines which solver from scipy.optimize is used, and it can be chosen from among the following strings: ?newton? for Newton-Raphson, ?nm? for Nelder-Mead ?bfgs? for Broyden-Fletcher-Goldfarb-Shanno (BFGS) ?lbfgs? for limited-memory BFGS with optional box constraints ?powell? for modified Powell?s method ?cg? for conjugate gradient ?ncg? for Newton-conjugate gradient ?basinhopping? for global basin-hopping solver The explicit arguments in fit are passed to the solver, with the exception of the basin-hopping solver. Each solver has several optional arguments that are not the same across solvers. See the notes section below (or scipy.optimize) for the available arguments and for the list of explicit arguments that the basin-hopping solver supports. maxiter : int, optional The maximum number of iterations to perform. full_output : bool, optional 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. disp : bool, optional Set to True to print convergence messages. fargs : tuple, optional Extra arguments passed to the likelihood function, i.e., loglike(x,*args) callback : callable callback(xk), optional Called after each iteration, as callback(xk), where xk is the current parameter vector. retall : bool, optional Set to True to return list of solutions at each iteration. Available in Results object?s mle_retvals attribute. skip_hessian : bool, optional If False (default), then the negative inverse hessian is calculated after the optimization. If True, then the hessian will not be calculated. However, it will be available in methods that use the hessian in the optimization (currently only with ?newton?). kwargs : keywords All kwargs are passed to the chosen solver with one exception. The following keyword controls what happens after the fit: 1 2 3 warn_convergence : bool, optional     If True, checks the model for the converged flag. If the     converged flag is False, a ConvergenceWarning is issued.

Notes

The ?basinhopping? solver ignores maxiter, retall, full_output explicit arguments.

Optional arguments for solvers (see returned Results.mle_settings):

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91

'newton'     tol : float         Relative error in params acceptable for convergence. 'nm' -- Nelder Mead     xtol : float         Relative error in params acceptable for convergence     ftol : float         Relative error in loglike(params) acceptable for         convergence     maxfun : int         Maximum number of function evaluations to make. 'bfgs'     gtol : float         Stop when norm of gradient is less than gtol.     norm : float         Order of norm (np.Inf is max, -np.Inf is min)     epsilon         If fprime is approximated, use this value for the step         size. Only relevant if LikelihoodModel.score is None. 'lbfgs'     m : int         This many terms are used for the Hessian approximation.     factr : float         A stop condition that is a variant of relative error.     pgtol : float         A stop condition that uses the projected gradient.     epsilon         If fprime is approximated, use this value for the step         size. Only relevant if LikelihoodModel.score is None.     maxfun : int         Maximum number of function evaluations to make.     bounds : sequence         (min, max) pairs for each element in x,         defining the bounds on that parameter.         Use None for one of min or max when there is no bound         in that direction. 'cg'     gtol : float         Stop when norm of gradient is less than gtol.     norm : float         Order of norm (np.Inf is max, -np.Inf is min)     epsilon : float         If fprime is approximated, use this value for the step         size. Can be scalar or vector.  Only relevant if         Likelihoodmodel.score is None. 'ncg'     fhess_p : callable f'(x,*args)         Function which computes the Hessian of f times an arbitrary         vector, p.  Should only be supplied if         LikelihoodModel.hessian is None.     avextol : float         Stop when the average relative error in the minimizer         falls below this amount.     epsilon : float or ndarray         If fhess is approximated, use this value for the step size.         Only relevant if Likelihoodmodel.hessian is None. 'powell'     xtol : float         Line-search error tolerance     ftol : float         Relative error in loglike(params) for acceptable for         convergence.     maxfun : int         Maximum number of function evaluations to make.     start_direc : ndarray         Initial direction set. 'basinhopping'     niter : integer         The number of basin hopping iterations.     niter_success : integer         Stop the run if the global minimum candidate remains the         same for this number of iterations.     T : float         The "temperature" parameter for the accept or reject         criterion. Higher "temperatures" mean that larger jumps         in function value will be accepted. For best results         `T` should be comparable to the separation (in function         value) between local minima.     stepsize : float         Initial step size for use in the random displacement.     interval : integer         The interval for how often to update the `stepsize`.     minimizer : dict         Extra keyword arguments to be passed to the minimizer         `scipy.optimize.minimize()`, for example 'method' - the         minimization method (e.g. 'L-BFGS-B'), or 'tol' - the         tolerance for termination. Other arguments are mapped from         explicit argument of `fit`:           - `args` <- `fargs`           - `jac` <- `score`           - `hess` <- `hess`