$Id: README,v 1.5 1998/03/16 08:21:26 kanungo Exp kanungo $ Package: UMDHMM version 1.02 Author: Tapas Kanungo (kanungo@cfar.umd.edu) Organization: University of Maryland, Collge Park, MD Web: http://www.cfar.umd.edu/~kanungo Date: 19 February, 1998 Updated on 5 May, 1999: see CHANGES file. Updated on 6 May, 1999: see CHANGES file. This software contains code for understanding the basics of hidden Markov models (HMM). The notation used is very similar to that used by Rabiner and Juang in: - Rabiner, L. R. and B. H. Juang, "Fundamentals of Speech Recognition," Prentice Hall, 1993. - Rabiner, L. R., "A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition, Prov. of IEEE, vol. 77, no. 2, pp. 257-286, 1989. - Rabiner, L. R., and B. H. Juang, "An Introduction to Hidden Markov Models," IEEE ASSP Magazine, vol. 3, no. 1, pp. 4-16, Jan. 1986. --------------------------------------------- Installation: --------------------------------------------- -------------------- UNIX: Dec, Sun Solaris, Linux (redhat): -------------------- Type "make all" at the unix prompt. It should compile the package. -------------------- Microsoft NT/95/98: -------------------- 1. Get the GNU package from: ftp://go.cygnus.com/pub/sourceware.cygnus.com/cygwin/latest/full.exe This package includes gcc and various commands and shells (sh, bash, etc.) that make the PC have a unix like environment. 2. Change to the UMDHMM directory and type "make all". --------------------------------------------- Executables: --------------------------------------------- genseq: Generates a symbol sequence using the specified model testvit: Generates the most like state sequence for a given symbol sequence, given the HMM, using Viterbi. esthmm: Estimates the HMM from a given symbol sequence using BaumWelch. testfor: Computes log Prob(observation|model) using the Forward algorithm. Note 1: The model test.hmm and sequence test.seq solve exercise 6.3 in the book by Rabiner and Juang (page 341). Just execute the command: prompt% testvit test.hmm test.seq and compare the output with the solution given in the book. --------------------------------------------- HMM file format: --------------------------------------------- M= <number of symbols> N= <number of states> A: a11 a12 ... a1N a21 a22 ... a2N . . . . . . . . . . . . aN1 aN2 ... aNN B: b11 b12 ... b1M b21 b22 ... b2M . . . . . . . . . . . . bN1 bN2 ... bNM pi: pi1 pi2 ... piN --------------------------------------------- Sample HMM file: --------------------------------------------- M= 2 N= 3 A: 0.333 0.333 0.333 0.333 0.333 0.333 0.333 0.333 0.333 B: 0.5 0.5 0.75 0.25 0.25 0.75 pi: 0.333 0.333 0.333 --------------------------------------------- Sequence file format: --------------------------------------------- T=<seqence lenght> o1 o2 o3 . . . oT --------------------------------------------- Sample sequence file: --------------------------------------------- T= 10 1 1 1 1 2 1 2 2 2 2