Hi, unfortunately this is not documented at all. The methods follows what Franz Och describes in his paper "Minimum Error Rate Training in Statistical Machine Translation".
-phi On 11/7/07, Ravishankar <[EMAIL PROTECTED]> wrote: > As part of my course work in NLP, I am working on statistical machine > translation using Moses software for English-Hindi translation. > I am currently working on rescoring strategy which could improve the > BLEU scores. For this i am using "mert-moses.pl" file by reusing some > parts of the code. > > In mert-moses.pl there are some sections of code which call a few > python scripts. I am familiar with perl but new to python. > I wanted how the scoring part is done. Is there any documentation > available as to how it is done. > > If any one has some documentation or can share some more information > about these, it would be helpful for me in my project work. > > Thanks, > Ravishankar > _______________________________________________ > Moses-support mailing list > [email protected] > http://mailman.mit.edu/mailman/listinfo/moses-support > > _______________________________________________ Moses-support mailing list [email protected] http://mailman.mit.edu/mailman/listinfo/moses-support
