[itsdb] parse big corpus with itsdb

[Francis Bond]

G'day,

>         However, I have another question, maybe someone can help me. I
> would like to use only one analysis per sentence, and I limited the number
> of answers to 1. But I don't know if this is the way to get the best
> possible analysis, is there some other switch that I could use?

in the grammar.set file (e.g. for the ERG pet/english.set) tel it to
read a scoring model:
sm := "vm6p.mem".
(see http://wiki.delph-in.net/moin/PetParameters).

I think this automatically turns ranking on.

You can then get only one result (the highest ranked) by calling cheap
with -nresults=1.

With a new pet and a new scoring model you can turn selective
unpacking on and get the best results with less memory by calling
cheap with:
-packing=15 -nsolutions=1
(see http://wiki.delph-in.net/moin/PetOptions)

-- 
Francis Bond  <www.kecl.ntt.co.jp/icl/mtg/members/bond/>
NTT Communication Science Laboratories | Natural Language Research Group