If you already have the CPE running, you can pass the descriptor to the
command line:
*org.apache.ctakes.ytex.tools.RunCPE or *
*org.apache.ctakes.core.cpe.CmdLineCpeRunner or*
*org.apache.uima.examples.cpe.SimpleRunCPE
A question about the default pipelines. There has been some concern
about the new assertion modules (the machine learning ones that I worked
on), partially due to some less intuitive error modes than negex and
partially due to its reliance on the dependency parser which increases
the memory
Sekhar,
There are a few open Jira's:
I think it would be a great contribution if you get this to work:
- CTAKES-189 https://issues.apache.org/jira/browse/CTAKES-189
GSoC: Implement OCR/Tika to standardize text input for cTAKES
-
- CTAKES-105
+1 for pushing forward
I may have been one of the voices commenting on memory bloat, but I agree with
Pei re: improving the new. The more use, the more attention and more
improvement (hopefully). I can't speak of the accuracy old v. new as I haven't
actually comparatively tested them. And
My vote would be to push forward.
The old assertion module also had it's share of bugs/issues and gives an
incentive to improve the new models.
And there's currently always the option for a user to easily revert back to the
old since it's not removed yet...
--Pei
-Original Message-
+ for push
Regards,
Taposh D. Roy | Health Data Project Lead/Scientist | Delivery System
Analytics, Decision Support | Kaiser Permanente | cell: 510.206.1633
|taposh.d@kp.org | 1950 Franklin Street, 17th Floor, Oakland, California
94588
NOTICE TO RECIPIENT: If you are not the
