Hi,

not quite sure what you are asking here.

You can use POS tags to assist language modelling
by adding a second "language model" over these
tags. But then, you also need the parallel corpus
to have POS tags on the target side (i.e., the translation
model has to generate them).

Also, if you have POS tagged corpus, you can train
a POS tagger with any number of toolkits, and then
tag all the data you have.

-phi

On Tue, May 29, 2012 at 2:03 PM, Ambaye Tadesse <[email protected]> wrote:
> I have data which is POS tagged and target side not tagged is there a
> problem if I use both of the data for Language model?
>
> Thanks
>
> _______________________________________________
> 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