Create the saddlepoint equality constraint function

This function creates and returns a function of the form function(a) {...}, where 'a' combines tvec and theta arguments to a single vector, in that order.

Usage

get.saddlepoint.eq.constraint.function(tvec, theta, observed.data, cgf)

Arguments

tvec

A numeric vector.

theta

A numeric vector.

observed.data

A numeric vector. See TO DO list.

cgf

An object of class 'CGF'.

Value

A function that accepts a vector 'a' as an argument. When a = c(tvec, theta) is passed to this function, it generates a list containing 'constraints' and 'jacobian'. 'constraints' are computed as \(K'(t;\theta) - y\), and 'jacobian' represents the gradient of these constraints with respect to both tvec and theta.

Details

In this set up, the expression of the saddlepoint equation is defined by \(K'(t;\theta) - y = 0\) for observed.data = y. The returned function is designed to compute both \(K'(t;\theta) - y\) and its gradient with respect to both tvec and theta.

Examples

if (FALSE) { # \dontrun{
# TO DO: write a working example
 # f <- get.saddlepoint_eq_constraint.function(tvec, theta, observed.data, cgf)
 # f(c(tvec, theta)) # returns a list of the form list(constraints = , jacobian = )
} # }