Convergence Test Based on L2 Norm

Share:

Description

Performs a test of convergence based on the L2 norm of the change in the parameter estimates.

Usage

1
conv.test(theta1, theta2, epsilon)

Arguments

theta1

vector of parameter estimates at previous step.

theta2

vector of parameter estimates at current step.

epsilon

positive convergence tolerance.

Details

This is used as the convergence test in the addreg fitting functions, because the EM algorithm may converge slowly such that the test based on the deviance used in glm.fit (see glm.control) may report convergence at a point away from the actual optimum.

Value

A logical; TRUE if sqrt(sum((theta1-theta2)**2))/sqrt(sum(theta1**2)) < epsilon, FALSE otherwise.

Author(s)

Mark W. Donoghoe Mark.Donoghoe@mq.edu.au

Examples

1
2
3
4
5
theta.old <- c(4,5,6)
theta.new <- c(4.05,5,6)

conv.test(theta.old, theta.new, 0.01)
conv.test(theta.old, theta.new, 0.005)