On the convergence of the iterative solution of the likelihood equations

R. Moddemeijer

University of Groningen, Department of Computing Science,
P.O. Box 800, NL-9700 AV Groningen, The Netherlands,
phone: +31.50.363 3940 - fax: +31.50.363 38005 - e-mail: rudy@cs.rug.nl

Abstract

To determine the maximum likelihood estimate in case of the iterative solution of the likelihood equations we need a convergence criterion. Usually numerical convergence criteria have been used; we propose a statistical convergence criterion. Our criterion reduces the number of iterations tremendously. The iterative algorithm to solve the equations is known to be unstable in the neighborhood of the exact solution. Our method avoids this numerical unstable region. The theory is validated by simulations.

Keywords

likelihood equations, convergence, maximum likelihood estimation, simulation, stability

Published

University of Groningen, Department of Computing Science, Groningen, Computing Science Report: CS-R9711, BibTeX
other publications