em.fit_pracma: Carry our the Newton-Raphson optimization of the parameters...

In emax.glm: General Tools for Building GLM Expectation-Maximization Models

Description

Carry our the Newton-Raphson optimization of the parameters for given weights via the pracma hessian,

Usage

em.fit_pracma(u, b, x, y, class_probs, weight, tol = 1e-08,
  debug = FALSE, family = poisson(), maxiter = Inf)

Arguments

u	A 'model.loglike' function.
b	The starting parameters.
x	An n-by-p design matrix.
y	A vector of observation of length n.
class_probs	An n length vector of probabilities for the proposed model.
weight	A n length vector of observation weight terms. This is currently designed to be either the exposure for a Poisson model or the number of trials for a Logistic model.
tol	The tolerance to repeat the Newton-Raphson optimization till.
debug	Debugging flag - set to TRUE to output step-by-step change in parameter values.
family	The GLM family being considered.
maxiter	Maximum number of NR steps to take.

Value

The parameter values on convergence.

Examples

x <- model.matrix(~ 1 + factor(wool) + factor(tension), data = warpbreaks)
y <- warpbreaks$breaks
class_probs = rep(1,54)
b <- c(1, 1, 1, 1)

u <- make.logLike(x, y, class_probs = class_probs)

em.fit_pracma(u, b, x, y, class_probs, weight = c(1))

emax.glm documentation built on July 4, 2019, 5:04 p.m.