Orosz György <oros...@itk.ppke.hu> writes: > Dear All, > > I am asking your help, hope someone can clarify these thigs: I am wondering > if it is > possible to use the apertium tagger as a standalone application, without > creating all > the resources used by the MT system.
It's possible to use it by itself, like echo '^foo/foo<n><sg>/foo<ij>$ ^bar/bar<n><sg>/bar<vblex><inf>$' | apertium-tagger en.prob I don't think you can compile apertium-tagger without compiling other core apertium functions, if that's what you mean. But the only other build dependency you have to compile is lttoolbox, and when you've got them installed, apertium-tagger is perfectly usable by itself (without using the rest of the scripts). > We have a morphologically disambiguated training > corpus, and a morphological analyzer. Is it possible to train the tagger in a > supervised > mode using only these resources? (Of course we can convert the output of the > MA to the > format that is used by Apertium.) > > If not, could anyone explain how to use the tool? > The manual states the following parameters are needed: > apertium-tagger[-d] -s=n DIC CRP TSX TAGGER_DATA HTAG UNTAG http://wiki.apertium.org/wiki/Tagger_training should be a good starting point (it seems no one has written the section on Supervised training, but see Question 2 under http://wiki.apertium.org/wiki/Unsupervised_tagger_training#Improving_the_tagger_performance ). hope this helps, Kevin ------------------------------------------------------------------------------ This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure _______________________________________________ Apertium-stuff mailing list Apertium-stuff@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/apertium-stuff
