As part of a MITRE internally funded R&D program in automatic speech
recognition, I have written a suite of programs for training ``Hidden
Markov Models'' using the ``Segmental K-Means Algorithm.'' This suite
includes shell scripts and programs written in C and Haskell.
A corrected version of the source code and some test data for the
Haskell programs (and only the Haskell programs) is now available by
anonymous ftp from
haskell.cs.yale.edu
in the directory
pub/haskell/progs
in the file
hmms-2.1.tar.gz
Last spring I released version 2.0 of this software. This version,
2.1, corrects two mistakes in the calculations and makes some other
minor changes.
The software has been thoroughly tested with hbc 0.999.4, and appears
to work with 0.999.7 (it does not work with 0.999.5). It also works
with ghc 0.22. It has not been tested with the Yale system.
Please note that I am no longer at MITRE. I will be happy to answer
questions about the software if I can, but because of a new job my
time will be limited.
David M. Goblirsch ([EMAIL PROTECTED])
NYNEX Science & Technology, Inc.
500 Westchester Ave.
White Plains, NY 10604
