Calculates and adds the scores to a fitted model object

Calculates and adds the scores to a fitted model object. The scores are the gradient observations/first derivatives of the observation-level log-likelihood function. The function is a wrapper around numDeriv::jacobian().

Usage

add_scores(object, func, x = coef(object), log_transform = is_bgw(object), ...)

Arguments

object

A fitted model object

func

A function with real (vector) results returning the function values at the observation level. This is either the log-likelihood function, or, as expected by bgw::bgw_mle(), the likelihood function (probabilities).

x

A real or real vector argument to func, indicating the point at which the gradient is to be calculated. Defaults to the estimated coefficients.

log_transform

A logical value indicating if func returns likelihoods that must be log transformed before differentiating. Defaults to TRUE for objects of class bgw_mle and FALSE otherwise.

...

Additional arguments passed to numDeriv::jacobian().

Value

A fitted model object with the scores added to the object.