rule-based negation annotator

["Miller, Timothy" <Ti...@childrens.harvard.edu>]

Hello,
I am starting to look at the negation annotators since we've had some reports that the machine learning module is missing some obvious negations. I've checked in an updated *Rule Based* version so that I can use it in modern uimafit-style pipelines. It has all the same default parameters so this shouldn't affect anyone using it "out of the box". I have decided not to spend time testing it in an old-style xml descriptor with modified parameters, because I think most people are (or should be) moving away from that style, and not using that annoator anyways. If you are using it that way I can't guarantee it will still work the same. Please let me know (via this list) if you run into any issues and we can work through it together.

Thanks
Tim

?

Re: rule-based negation annotator

["Miller, Timothy" <Ti...@childrens.harvard.edu>]

Yes, it's just an API upgrade to the negex-approximator included in
ctakes-ne-contexts. If you are using it and want to give it a test I
would be very grateful!
Tim


On 01/06/2016 01:36 PM, David Kincaid wrote:
> Is this new negation module different that the NE Contexts negation
> annotator already in cTAKES or is it an update of it? Is it still based on
> NegEx?
>
> - Dave
>
> On Wed, Jan 6, 2016 at 11:31 AM, Miller, Timothy <
> Timothy.Miller@childrens.harvard.edu> wrote:
>
>> Hello,
>> I am starting to look at the negation annotators since we've had some
>> reports that the machine learning module is missing some obvious negations.
>> I've checked in an updated *Rule Based* version so that I can use it in
>> modern uimafit-style pipelines. It has all the same default parameters so
>> this shouldn't affect anyone using it "out of the box". I have decided not
>> to spend time testing it in an old-style xml descriptor with modified
>> parameters, because I think most people are (or should be) moving away from
>> that style, and not using that annoator anyways. If you are using it that
>> way I can't guarantee it will still work the same. Please let me know (via
>> this list) if you run into any issues and we can work through it together.
>>
>> Thanks
>> Tim
>>
>> ?
>>
>>

Re: rule-based negation annotator

[David Kincaid <ki...@gmail.com>]

Is this new negation module different that the NE Contexts negation
annotator already in cTAKES or is it an update of it? Is it still based on
NegEx?

- Dave

On Wed, Jan 6, 2016 at 11:31 AM, Miller, Timothy <
Timothy.Miller@childrens.harvard.edu> wrote:

> Hello,
> I am starting to look at the negation annotators since we've had some
> reports that the machine learning module is missing some obvious negations.
> I've checked in an updated *Rule Based* version so that I can use it in
> modern uimafit-style pipelines. It has all the same default parameters so
> this shouldn't affect anyone using it "out of the box". I have decided not
> to spend time testing it in an old-style xml descriptor with modified
> parameters, because I think most people are (or should be) moving away from
> that style, and not using that annoator anyways. If you are using it that
> way I can't guarantee it will still work the same. Please let me know (via
> this list) if you run into any issues and we can work through it together.
>
> Thanks
> Tim
>
> ?
>
>