Penalized maximum-likelihood estimation, the Baum-Welch algorithm, diagonal balancing of symmetric matrices and applications to training acoustic data

Charles A. Micchelli; Peder Olsen

doi:10.1016/S0377-0427(00)00385-X

J. Comput. Appl. Math.

Paper

01 Jul 2000

Penalized maximum-likelihood estimation, the Baum-Welch algorithm, diagonal balancing of symmetric matrices and applications to training acoustic data

View publication

Abstract

We study penalized maximum-likelihood estimation methods for nonparametric density estimation and propose their use in training acoustic models for speech recognition. Several algorithms for the numerical solution of the optimization problems that we encounter are proposed and analyzed. © 2000 Elsevier Science B.V.

Conference paper