[BioNLP] (no subject)

Bob Futrelle bob.futrelle at gmail.com
Tue May 22 10:51:08 EDT 2012

Works for me.  I pasted the "Cell sample" (365 words) into a curl command
specifying json output and a redirected the output to a file.  It worked
quickly. They are to be commended for what they have accomplished in Cocoa.
They have taken on quite a challenge, though some results are questionable
- no surprise.  Also, the tag set is small. Easy to imagine post-processing
to refine the tags.

For example, the bigram  "antiviral efficacy" comes out as

        "_type": "Physiology",
        "_typeGroup": "entities",
        "name": "antiviral",
        "instances": [
                "exact": "antiviral",
                "offset": 2194,
                "length": 9
        "_type": "Process",
        "_typeGroup": "entities",
        "name": "efficacy",
        "instances": [
                "exact": "efficacy",
                "offset": 2204,
                "length": 8

with "efficacy" tagged as type "Process".  The pair "antiviral efficacy"
describes a certain degree of effectiveness of the prior-mentioned
"inhibitory siRNAs".  [No one said natural language would be easy.]

All in all a great tool which can be used for exploration or for more
extended tasks. I would worry that someone would write a script that would
send a few thousand curl lines.  So proceed cautiously.

I'm busy developing NLP tools of my own (in deep cover), so I don't really
use any other tools.  Not sure what else out there may have functionality
similar to Cocoa. I'm sure some list members will mention similar tools.

 - Bob Futrelle

On Tue, May 22, 2012 at 7:35 AM, S. V. Ramanan <ramanan.sv at gmail.com> wrote:

> Dear BioNLP members,
>      We are pleased to announce Cocoa, a dense annotator for
> biological text. Cocoa annotates macromolecules, chemicals, protein/
> DNA parts, organisms, processes, anatomical parts, locations,
> physiological terms, parameters, values, experimental techniques, and
> foods. In addition to a web frontend, Cocoa is also accessible
> through a WebAPI, which returns markup in JSON and A1 (GENIA/Brat)
> formats.
> Availability: http://npjoint.com
> We welcome feedback and comments (contact at npjoint.com) from the
> BioNLP community on any aspect of our product.
> Thank you for your time.
> - S. V. Ramanan; ramanan at npjoint.com
> _______________________________________________
> BioNLP mailing list
> BioNLP at lists.ccs.neu.edu
> https://lists.ccs.neu.edu/bin/listinfo/bionlp
> The BioNLP website: http://www.bionlp.org
-------------- next part --------------
HTML attachment scrubbed and removed

More information about the BioNLP mailing list