regmixPhiStep: regmixPhiStep

In hkasahar/normalregMix: normalregMix: An R Package for Normal Mixture Regression Models

Description

Given a pair of h and tau and data, compute ordinary & penalized log-likelihood ratio resulting from MEM algorithm at k=1,2,3, tailored for parallelization.

Usage

regmixPhiStep(htaupair, y, x, parlist, z = NULL, p, an, ninits, ninits.short,
  epsilon.short, epsilon, maxit.short, maxit, verb)

Arguments

htaupair	A set of h and tau
y	n by 1 vector of data for y
x	n by q matrix of data for x
parlist	The parameter estimates as a list containing alpha, mu, sigma, and gam in the form of (alpha = (alpha_1, ..., alpha_m), mu = (mu_1, ..., mu_m), sigma = (sigma_1, ..., sigma_m), gam = (gam_1, ..., gam_m))
z	n by p matrix of regressor associated with gam
p	Dimension of z
an	a term used for penalty function
ninits	The number of randomly drawn initial values.
ninits.short	The number of candidates used to generate an initial phi, in short MEM
epsilon.short	The convergence criterion in short EM. Convergence is declared when the penalized log-likelihood increases by less than epsilon.short.
epsilon	The convergence criterion. Convergence is declared when the penalized log-likelihood increases by less than epsilon.
maxit.short	The maximum number of iterations in short EM.
maxit	The maximum number of iterations.
verb	Determines whether to print a message if an error occurs.

Value

A list of coefficients, log-likelihood, and penalized log-likelihood resulting from MEM algorithm.

hkasahar/normalregMix documentation built on May 17, 2019, 4 p.m.