search

biplab-iitb / practNLPTools

Practical Natural Language Processing Tools for Humans. Dependency Parsing, Syntactic Constituent Parsing, Semantic Role Labeling, Named Entity Recognisation, Shallow chunking, Part of Speech Tagging, all in Python.
192 stars 61 forks source link

interpretation of the srl result #2

Open aleissaf opened 8 years ago

aleissaf commented 8 years ago

Dear,

I have been using the tool for semantic role labelling, but I haven't been able to understand the result of the semantic role labelling. Can someone please guide me. Text: RT @WBBMNewsradio: Gov. Rauner Sends National Guard To Southern Illinois In Response To Flooding - Full Story: Result [{'A1': 'National Guard', 'A0': 'Gov. Rauner', 'A2': 'To Southern Illinois', 'AM-PNC': 'In Response To Flooding - Full Story', 'V': 'Sends'}]

Thanks,

profversaggi commented 8 years ago

We're using it successfully for one of out projects. I've been able to gather a good bit of that information from a different source but it might be useful....let me know if you still need it.

On Mon, Feb 22, 2016 at 3:36 PM, aleissaf notifications@github.com wrote:

Dear,

I have been using the tool for semantic role labelling, but I haven't been able to understand the result of the semantic role labelling. Can someone please guide me. Text: RT @WBBMNewsradio: Gov. Rauner Sends National Guard To Southern Illinois In Response To Flooding - Full Story: Result [{'A1': 'National Guard', 'A0': 'Gov. Rauner', 'A2': 'To Southern Illinois', 'AM-PNC': 'In Response To Flooding - Full Story', 'V': 'Sends'}]

Thanks,

— Reply to this email directly or view it on GitHub https://github.com/biplab-iitb/practNLPTools/issues/2.

######################################################### Matthew R. Versaggi, Artificial Intelligence Engineer, Imagine One, LTD President & CEO: Versaggi Information Systems, Inc. Adjunct Professor of eBusiness DePaul University Email: mailto:matt@versaggi.com, ProfVersaggi@gmail.com M: 630-292-8422 LinkedIn: http://www.linkedin.com/in/versaggi About Me: http://www.matt-versaggi.com/resume/ #########################################################

aleissaf commented 8 years ago

Dear,

As there any way that I can extract only the events from the semantic role labelling?

Thanks,

profversaggi commented 8 years ago

Yes, we've done specifically that in our project. I have extensive notes on exactly how that is done. We used SRL and PyKE (inference engine) to accomplish that. It's a fusion of technologies but it works.

If you like I can forward what I have to you in order to walk you through it ...

On Fri, Feb 26, 2016 at 5:58 PM, aleissaf notifications@github.com wrote:

Dear,

As there any way that I can extract only the events from the semantic role labelling?

Thanks,

— Reply to this email directly or view it on GitHub https://github.com/biplab-iitb/practNLPTools/issues/2#issuecomment-189517361 .

######################################################### Matthew R. Versaggi, Artificial Intelligence Engineer, Imagine One, LTD President & CEO: Versaggi Information Systems, Inc. Adjunct Professor of eBusiness DePaul University Email: mailto:matt@versaggi.com, ProfVersaggi@gmail.com M: 630-292-8422 LinkedIn: http://www.linkedin.com/in/versaggi About Me: http://www.matt-versaggi.com/resume/ #########################################################

aleissaf commented 8 years ago

Dear Versaggi,

I would really appreciate it if you can share anything that guides me through the process.

Thank you very much

profversaggi commented 8 years ago

I'd be glad to, give me an idea of what it is you are doing and the context and I'll see if I can cater it better to what you are doing....

On Sun, Feb 28, 2016 at 3:13 PM, aleissaf notifications@github.com wrote:

Dear Versaggi,

I would really appreciate it if you can share anything that guides me through the process.

Thank you very much

— Reply to this email directly or view it on GitHub https://github.com/biplab-iitb/practNLPTools/issues/2#issuecomment-189936620 .

######################################################### Matthew R. Versaggi, Artificial Intelligence Engineer, Imagine One, LTD President & CEO: Versaggi Information Systems, Inc. Adjunct Professor of eBusiness DePaul University Email: mailto:matt@versaggi.com, ProfVersaggi@gmail.com M: 630-292-8422 LinkedIn: http://www.linkedin.com/in/versaggi About Me: http://www.matt-versaggi.com/resume/ #########################################################

aleissaf commented 8 years ago

Basically, i'm trying to read and analyze twitter news feed. I want to identify from the results only the events after removing the stop words.

profversaggi commented 8 years ago

Hi Aleissaf,

I used a different approach. I leveraged an Inference engine to pattern match only those patterns which would typically correspond to an event. That worked well for me ...

I've included a process flow diagram which shows how I did it, and can give you code snippets of that'll help.

In our case we had different requirements, we were picking out only specific events in a news feed so we only wanted those events whose verb corresponded to a DB of (selected) verbs which we seeded (and would be indicative of the kinds of events we were tracking.

In most cases that worked really well for us to capture the universe of events and then allow us to whittle them down to just what we wanted.

Let me know if this helps and if you need some code to peruse.

-matt

On Mon, Feb 29, 2016 at 7:50 AM, aleissaf notifications@github.com wrote:

Basically, i'm trying to read and analyze twitter news feed. I want to identify from the results only the events after removing the stop words.

— Reply to this email directly or view it on GitHub https://github.com/biplab-iitb/practNLPTools/issues/2#issuecomment-190196952 .

######################################################### Matthew R. Versaggi, Artificial Intelligence Engineer, Imagine One, LTD President & CEO: Versaggi Information Systems, Inc. Adjunct Professor of eBusiness DePaul University Email: mailto:matt@versaggi.com, ProfVersaggi@gmail.com M: 630-292-8422 LinkedIn: http://www.linkedin.com/in/versaggi About Me: http://www.matt-versaggi.com/resume/ #########################################################

profversaggi commented 8 years ago

I don't think that the process flow diagram came through via the email response, so I added it here...

srl-pyke-process-flow

subhadeepmaji commented 8 years ago

Please refer to this for a detailed description of Propbank SRL labels. https://verbs.colorado.edu/~mpalmer/projects/ace/PBguidelines.pdf

SENNA uses Propbank labeling, other popular choice is FrameNet.