Hi Mehnaz,

It seems that the problem is the same in both cases (He and Mg). As these are very common words in speech, they tend to bring much more noise, than useful information. It is best if you remove them from the knowledge base. Use intelligent Jape rules instead to take into consideration the context in which these words appear. Then create annotations in the RHS of the jape rules. For example for the "mg" to not match weight measure units , you could try this rule:

Rule: MgRule
Priority: 100
(
   ({Token.kind != "number"})
   ({Token.string == "mg}):mg
)
-->
{
 // Create the annotation here
}


All the best,
Philip



Mehnaz Adnan wrote:
Hi
Here is the answers *- what is the text you are running the extraction on
*
The example text is
" He is using 1mg paradol" *- what pipeline you are executing
*
At this time I am using default KIM pipline (for testing, but need to define my own in future). I have replaced wkb.nt with my own wkb.nt which have medical terminiogies and their aliases in which "he" "mg" and "paradol" is included. I ran toolpopulate.cmd on my test file and in KIM web UI it creates three annotations like this: He is using 1mg paradol I just want to create annotation for paradol but not for He and mg (b/c he has a diffrenent meanign as this He ). Please guide me how can I do this . I am using Kimplatform-2.3-RC1 Thanks Mehnaz

2009/3/25 Philip Alexiev <[email protected] <mailto:[email protected]>>

    Hi Mehnaz,

    Please provide more context to help us understand the problem.

    - what is the text you are running the extraction on
    - what pipeline you are executing
    - what version of KIM you are using

    Any additional information will be useful.

    Greetings,
    Philip


    Mehnaz Adnan wrote:

        Hi
         I want to treat "1mg" as a single word for annotation.
        Currently an annotation is created for "mg" b/c "mg" is a
        resource in my knowledge badse. I think for this I have to
        make some changes in tokenization rule.I found three jape
        grammars in kim. I am not sure which one the kim is using for
        creating tokens.
          Next question where I found some prelliminary information to
        make changes in IE module of KIM.
         Thanks in advance

-- Mehnaz Adnan
        Ph.D. Candidate,
        Department of Computer Science-Tamaki
        University of Auckland
        Phone: 09-3737599 ext 83274
        email: [email protected]
        <mailto:[email protected]>
        <mailto:[email protected]
        <mailto:[email protected]>>
        ------------------------------------------------------------------------

        _______________________________________________
        Kim-discussion mailing list
        [email protected] <mailto:[email protected]>
        http://ontotext.com/mailman/listinfo/kim-discussion




--
Mehnaz Adnan
Ph.D. Candidate,
Department of Computer Science-Tamaki
University of Auckland
Phone: 09-3737599 ext 83274
email: [email protected] <mailto:[email protected]>

_______________________________________________
Kim-discussion mailing list
[email protected]
http://ontotext.com/mailman/listinfo/kim-discussion

Reply via email to