ctakes-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Murali Nagendranath <mmin...@gmail.com>
Subject Re: spell corrector in cTakes
Date Tue, 17 Nov 2015 14:16:06 GMT
Hello Ashutosh,

As the new module is based on ML, the need for training it on large data
set is a key factor. I do not have a timeline to share with you. We are
definitely considering the use case that you have indicated as part of our
test cases.

Best,
Murali

On Mon, Nov 16, 2015 at 6:06 PM, Ashutosh Modi <modiashutosh@gmail.com>
wrote:

> Hi Murali,
>
> Thanks for the info! When is it expected to be released? Is it possible to
> get a chance to beta test it? Also can it handle (may be split) things like
> words which have been joined into one for e.g. lymphnode (lymph+node),
> because in such cases, cTakes fails to identify such UMLS entities.
>
> Best,
> Ashutosh
>
> On Mon, Nov 16, 2015 at 5:59 PM, Murali <mminnah@gmail.com> wrote:
>
>> Ashutosh
>>
>> Great idea.
>>
>> Wired Informatics has built a spell corrector and is looking forward to
>> contributing back to cTAKES once it's fully tested.
>> There is something in sandbox, but it hasn't been vetted out nor has the
>> full training corpus yet.
>>
>> Best
>> Murali
>>
>> > On Nov 16, 2015, at 5:40 PM, Ashutosh Modi <modiashutosh@gmail.com>
>> wrote:
>> >
>> > Hi,
>> >
>> > I was wondering if there is some spelling correction component
>> available in cTakes. Ideally, the component would makes use of both the
>> english lexicon as well the lexicons in UMLS.
>> >
>> > In case such a component exists, please let me know.
>> >
>> > Thanks,
>> > Ashutosh
>>
>
>

Mime
View raw message