Hi Michael.
The number you are asking for is the third one, "1" in your example.
The five numbers are, in order:
Phrase translation probability of the source phrase given the target phrase
Lexical weighting of the source phrase given the target phrase
Phrase translation of the target phrase given the source phrase (the one you
are asking for)
Lexical weighting of the target phrase given the source phrase
Phrase penalty (always exp(1) = 2.718)
For more information you may want to give a look at the Moses web page, which
has a section about "Scoring Phrases"
(http://www.statmt.org/moses/?n=FactoredTraining.ScorePhrases).
Bye.
--
Carlos A. Henríquez Q.
+34-693-278-219
[EMAIL PROTECTED]
[EMAIL PROTECTED]
----- Mensaje original ----
De: Michael Zuckerman <[EMAIL PROTECTED]>
Para: [email protected]
Enviado: martes, 8 de julio, 2008 15:07:51
Asunto: [Moses-support] Statistical likelihood of phrases in a phrase table
Hi,
In the phrase table that we get after training, we have lines, like:
! " in a crowded theater ||| " en un teatro ||| () (0) (1) (2) (3) (3) ||| (1)
(2) (3) (4,5) ||| 0.5 5.05228e-09 1 0.0370417 2.718
I think that one of the last 5 numbers of the line means the statistical
likelihood of a translation, i.e. how it is likely that "\" en un teatro" is a
translation of "! \" in a crowded theater". Could you please tell me what
number is this likelihood ?
Thank you very much,
Michael.
______________________________________________
Enviado desde Correo Yahoo! La bandeja de entrada más inteligente._______________________________________________
Moses-support mailing list
[email protected]
http://mailman.mit.edu/mailman/listinfo/moses-support
