Hi Ant B & James,
yes it is indeed strange...no I did not use a clone of your project -
using my own Clojure code, however since we use identical data and
models I'd expect identical results as well...Initially I was sure that
something was going wrong when inserting the named-entities into the
parse tree...that is not the case though - I confirmed it...your code
and my code insert exactly the same entities so it is only my coref code
that either 'does things differently' or we could actually be using
different opennlp versions...can I assume you're on the latest trunk or
are you still using the problematic 1.5.2 incubator version? I don't
know how much the coref component has changed since 1.5.2 but there are
several other bug-fxes that may interact with it (tokenizer, name-finder
& parser)...
as far as sharing snippets online is concerned, I'm perfectly fine with
that...the question is how much would you be able to help me with my
Clojure code? :-)
*@ James,*
as you 've probably gathered, we weren't able to resolve the coref
discrepancies between the 2 of us...did you have any luck running the
coref component on the standard example sentence, using trunk and the
latest wordnet, as originally requested? Doing this will help us
determine which one of the outputs we are comparing is more appropriate.
Ant B's output certainly *seems* more correct (from an NLP persective)
but I'm not giving up unless I find out *why* it is more correct and
*why* I'm not getting the same...I totally understand that you are
probably very very busy these days, with the upcoming 1.5.3 release and
all, but wouldn't it be nice to know that the coref component has not
been affected after all the fixes? presumably you must have tests in
place...
many many thanks,
Jim
On 09/03/13 18:50, Ant B wrote:
Hi Jim,
Sorry for the slow response. Strange that you got different results - do you
get different results when you use a clone of the project?
Sadly I don't have any advice to offer, but if you are using your own code, I'd
be happy to offer a second pair of eyes if you can share code snippets online?
Ant
On Mar 6, 2013, at 6:15 AM, Jim - FooBar(); <[email protected]> wrote:
right, finally the moment of truth! here is what I get using your loop:
Mention set:: [ this British industrial conglomerate ]
Mention set:: [ a director ]
Mention set:: [ Consolidated Gold Fields PLC ]
Mention set:: [ chairman :: former chairman ]
Mention set:: [ 55 years ]
Mention set:: [ Rudolph Agnew ]
Mention set:: [ Elsevier N . V . :: the Dutch publishing group ]
Mention set:: [ Mr . Vinken ]
Mention set:: [ a nonexecutive director Nov . 29 ]
Mention set:: [ the board ]
Mention set:: [ 61 years ]
Mention set:: [ Pierre Vinken ]
As you can see I am missing these 2 which do seem correct (in your output):
Mention set:: [ Pierre Vinken :: Mr. Vinken ]
Mention set:: [ a nonexecutive director :: chairman :: former chairman :: a
director ]
Jim
ps: I can confirm that the NEs can be retrieved correctly from the parse-tree
just like yours...
On 06/03/13 12:19, Jim - FooBar(); wrote:
ok found your code and it answers all my questions!....I'll do the same now and
see what happens... :)
Jim
