i think what you're doing is called 'sentence-level bleu'. google it to
see what other people are doing.
it's an approximation of the document bleu, but the exact number won't
be the same
On 22/11/2012 14:17, Cuong Hoang wrote:
Hi all,
I have a very naive question, as I guess so.
I'm currently try to write an BLEU score measurement program.
I just have a very naive wonder when I don't calculate BLEU score for
a *FILE, *which contains from 1,000 to 2,000 sentences,
in the traditional way. I calculate the BLEU score for each one,
and assign the BLEU score for the test File by the average
value of every pair.
As the result, I will obtain a quite low result in compare to the
traditional way.
So, which way is better ?
Thanks and best regards,
C. Hoang
--
Hoàng Cu+o+`ng
SMTNerd
_______________________________________________
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
