nbest-mix.1

这是一款很好用的工具包

nbest-mix(1)                                         nbest-mix(1)



NNAAMMEE
       nbest-mix - interpolate N-best posterior probabilities

SSYYNNOOPPSSIISS
       nnbbeesstt--mmiixx  [--hheellpp]  option  ...   _w_e_i_g_h_t_1  _n_b_e_s_t_1  _w_e_i_g_h_t_2
       _n_b_e_s_t_2 ...

DDEESSCCRRIIPPTTIIOONN
       nnbbeesstt--mmiixx reads a number of N-best lists (which must  con-
       tain identical hypotheses), computes the hypothesis poste-
       rior probabilities for each, and computes a new  posterior
       distribution  that is a weighted mixture of the input dis-
       tributions.  The hypothesis with the highest combined pos-
       terior probability is printed.

       The  command  line  arguments  form an alternating list of
       weight values and N-best file names.

OOPPTTIIOONNSS
       Each filename argument can be an ASCII  file,  or  a  com-
       pressed file (name ending in .Z or .gz), or ``-'' to indi-
       cate stdin/stdout.

       --hheellpp  Print option summary.

       --vveerrssiioonn
              Print version information.

       --ddeebbuugg _l_e_v_e_l
              Controls the  amount  of  output  (the  higher  the
              _l_e_v_e_l, the more).

       --wwrriittee--nnbbeesstt _f_i_l_e
              Output  N-best  lists containing scores that corre-
              spond to the log of the combined posteriors of  the
              input hypotheses.  The log posterior is assigned as
              the acoustic score and  other  scores  are  set  to
              zero.   This  also  suppresses  the printing of the
              best hyp.

       --mmaaxx--nnbbeesstt _n
              Limits the number of hypotheses read from  each  N-
              best list to the first _n.

       --rreessccoorree--llmmww _l_m_w
              Sets  the  language  model weight used in combining
              the language model log probabilities with  acoustic
              log probabilities (only relevant if separate scores
              are given in the N-best input).

       --rreessccoorree--wwttww _w_t_w
              Sets the word transition weight used to weight  the
              number of words relative to the acoustic log proba-
              bilities (only  relevant  if  separate  scores  are
              given in the N-best input).

       --ppoosstteerriioorr--ssccaallee _s_c_a_l_e
              Divide  the  total weighted log score by _s_c_a_l_e when
              computing normalized posterior probabilities.  This
              controls  the peakedness of the posterior distribu-
              tion.  The default value is whatever was chosen for
              _l_m_w,  so  that  language model scores are scaled to
              have weight 1,  and  acoustic  scores  have  weight
              1/_l_m_w.

       --sseett--llmm--ssccoorreess
              In  conjunction  with  --wwrriittee--nnbbeesstt,  output N-best
              lists that preserve the acoustic  scores  and  word
              counts  of  the  (first of the) input N-best lists,
              and encodes the combined log posteriors via the  LM
              scores.  The LM scores in the output are calculated
              so that, when combined with the acoustic scores and
              insertion  penalties (using the given LM weight and
              posterior scaling), the  result  is  the  weighted,
              combined  posteriors  based  on  all  input  N-best
              scores.  This option  is  useful  if  input  N-best
              lists were created by rescoring with different lan-
              guage models, and the output N-best lists are to be
              combined  with other scores or if the score weight-
              ing is to be optimized with nnbbeesstt--ooppttiimmiizzee(1).

       --sseett--aamm--ssccoorreess
              Analogous to --sseett--llmm--ssccoorreess, except that the acous-
              tic  scores  are  modified  to reflect combined log
              posterior probabiltities, and other scores are pre-
              served.   This  option  is  useful  if input N-best
              lists were  created  by  rescoring  with  different
              acoustic models.

SSEEEE AALLSSOO
       nbest-lattice(1), nbest-scripts(1), nbest-optimize(1).
       A. Stolcke, K. Ries, N. Coccaro, E. Shriberg, R. Bates, D.
       Jurafsky, P. Taylor, R. Martin, C. Van  Ess-Dykema,  &  M.
       Meteer,  ``Dialogue Act Modeling for Automatic Tagging and
       Recognition of Conversational Speech,'' _C_o_m_p_u_t_a_t_i_o_n_a_l _L_i_n_-
       _g_u_i_s_t_i_c_s 26(3), 339-373, 2000.

BBUUGGSS
       Hopefully not.

AAUUTTHHOORR
       Andreas Stolcke <[e-mail]>.
       Copyright 1998-2004 SRI International



SRILM Tools        $Date: 2004/12/03 17:59:01 $      nbest-mix(1)