search

opener-project / pos-tagger-en-es

POS tagger for English, Spanish, Dutch and other languages.
Other
5 stars 1 forks source link

POS Tagger Personalization #1

Open elvissuarez opened 9 years ago

elvissuarez commented 9 years ago

Hi there

Let me know how could i modify the code for town recognition? because i am in colombia and the entity recognition doesn't work well with us towns.

Is it possible? Can POSTagger be modify in this way or what should I do?

thanks so much

yorickpeterse commented 9 years ago

For this to work you'll need to build your own models that can be used with the POS tagger. Some information on this can be found at https://github.com/opener-project/pos-tagger-en-es/tree/master/core#overview. How these specific models are composed is something I'm not really familiar with sadly.

elvissuarez commented 9 years ago

Hi There Build own model is only way to add towns to entity recognition or POStagging as locations.

thanks so much

angelo337 commented 9 years ago

hi there Is it possible to include more names of locations, and organizations in this POS tagger?. if so could you please point me out some resources to do the changes, and in which files, because as I see in OPENNLP, they have different models for locations organizations in every language but not in the OPENNER project, i don't see the same structure. thanks

yorickpeterse commented 9 years ago

@ragerri I'm not particular familiar with building the models, do we have any public documentation on this matter?

ragerri commented 9 years ago

Hi,

To annotate names of places, persons and so on you need to use the her component, not the pos tagger. The ner component is trainable, you can read about it here

https://github.com/opener-project/ner-base/tree/master/core#training

I am on holidays until the 27th, I can give more support after that.

Best,

R

On Tue, Aug 4, 2015 at 12:24 AM, Yorick Peterse notifications@github.com wrote:

@ragerri https://github.com/ragerri I'm not particular familiar with building the models, do we have any public documentation on this matter?

— Reply to this email directly or view it on GitHub https://github.com/opener-project/pos-tagger-en-es/issues/1#issuecomment-127430318 .

elvissuarez commented 9 years ago

@ragerri welcome back, would you help us with NER personalization? we are expecting that, thank you so much.

ragerri commented 9 years ago

Hello,

Did you read the ixa-pipe-nerc documentation about training new models? You need a corpus manually annotated with entities you are interested in and formatted in opennlp format. http://opennlp.apache.org/documentation/1.6.0/manual/opennlp.html#tools.namefind

Did you try to do it?