Hi,

the script implements IBM BLEU. The difference between IBM BLEU
and NIST BLEU is that when using multiple reference translations,
IBM BLEU uses the closest reference translation length (per sentece)
and NIST the shortest.

-phi

On Thu, Mar 26, 2009 at 8:45 AM, Joerg Tiedemann <[email protected]> wrote:
>
> hi,
>
>
> I was asked whether multi-bleu.perl gives IBM or NIST BLEU scores. I
> didn't know what to answer ...could someone tell me?
>
> thanks!
> --
>
> Jörg
>
>
> ***********/\/\/\/\/\/\/\/\/\/\/\************************************
> **  Jörg Tiedemann                 [email protected]              **
> **  Alfa-Informatica               http://www.let.rug.nl/~tiedeman **
> **  Rijksuniversiteit Groningen    Harmoniegebouw, room 1311-429   **
> **  Postbus 716                    phone: +31 (0)50-363 5935       **
> **  9700 AS Groningen              fax:   +31 (0)50-363 6855       **
> *************************************/\/\/\/\/\/\/\/\/\/\/\**********
>
> _______________________________________________
> 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

Reply via email to