ProbitUpdate: EM Algorithm Update for Probit Regression

In daarem: Damped Anderson Acceleration with Epsilon Monotonicity for Accelerating EM-Like Monotone Algorithms

EM Algorithm Update for Probit Regression

Description

Function performs an EM update (both the E and M steps) of the parameters for a Probit regression model.

Usage

ProbitUpdate(beta.hat, X, y)

Arguments

beta.hat	A vector of length p. The current estimates of the regression parameters.
X	The n x p design matrix for the Probit regression model.
y	Vector of length n containing binary outcomes (either 0 or 1).

Value

A vector of length p - the updated parameter values.

Author(s)

Nicholas Henderson

See Also

ProbitSimulate, ProbitLogLik

Examples

n <- 200
npars <- 5
true.beta <- .5*rt(npars, df=2) + 2
XX <- matrix(rnorm(n*npars), nrow=n, ncol=npars)
yy <- ProbitSimulate(true.beta, XX)

initial.beta <- rep(0.0, npars)
new.beta <- ProbitUpdate(initial.beta, XX, yy)

daarem documentation built on March 23, 2022, 9:06 a.m.