The tag dictionary is an POS Tagger thing, it can use it to lookup
the possible tags for a certain token. This increases the accuracy
and speeds up the tag sequence search.
We do not have support for such rules in the learn able
Name Finder.
There is a dictionary based name finder which can do this
and a regex based name finder.
If you want to stick to the learn able name finder you can implement
a custom sequence validator to force it to detect these tokens always
as names.
Do you train yourself? Then you could add more training data which contains
these names.
Jörn
On 07/13/2012 02:35 PM, Daniel wrote:
hmm perhaps I needn't use it, what I would like to do is to have a
dictionary or something where I put some words like it:
<start:person>Daniel</start>
<start:person>William</start>
<start:organisation>IBM</start>
<start:organisation>HP</start>
and then NameFinder always detect that tokens (Daniel, William, IBM, HP) as
person or organisation entities...is that possible?
Thanks!
and OpenNLP NameFinder recognize
2012/7/13 William Colen <[email protected]>
Hi, Daniel,
You can create one with the default implementation, that would serialize
your dictionary to a XML file, through the API:
http://svn.apache.org/viewvc/opennlp/trunk/opennlp-tools/src/main/java/opennlp/tools/postag/POSDictionary.java?view=markup
There is some usage example here:
http://svn.apache.org/viewvc/opennlp/trunk/opennlp-tools/src/test/java/opennlp/tools/postag/POSDictionaryTest.java?view=markup
Regards,
William
On Fri, Jul 13, 2012 at 9:08 AM, Daniel <[email protected]> wrote:
I have checked OpenNLP documentation, but there is very little
information
about Tag Dictionary. Can anyone tell me how create a dictionary and how
use it on one project?
Thanks!
