Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[java-nlp-user] Can Stanford NER-Jan 2009 be used for training of non English NER system like Hindi

John Bauer horatio at gmail.com
Fri Apr 8 17:17:07 PDT 2011


Sorry for the long delay in responding... it would be very hard to
diagnose from here without seeing what is going on.  If this is still
a problem, would you send some data files and a description on how you
produced the model?

Unless the training set and test set were completely disjoint, and you
used a small training set, it shouldn't be the case that nothing is
tagged at all in the test set,

John

On Thu, Dec 23, 2010 at 12:54 AM, ASHUTOSH SINGH
<singhashutosh90 at gmail.com> wrote:
> Respected sir,
>
> Yes sir, that is exactly what I am doing and it is generating a .ser.gz
> model also after the training. Then I am using this model for tagging of a
> new test file by the command java -cp stanford-ner.jar
> edu.stanford.nlp.ie.crf.CRFClassifier -loadClassifier myner-model.ser.gz
> -testFile my_test_file.txt  as given on your website. But then Stanford NER
> is tagging all the tokens with O in the test file.
> I am also attaching the model generated with the mail.
>
> With regards
> Ashutosh
> B.Tech 3rd year
> Department of Computer Science and Engineering
> NIT Tiruchirappalli
> India
>
>
>



More information about the java-nlp-user mailing list