Named entity extraction/correlation using Semantic Knowledge Graph

[Pratik Patel <pr...@semandex.net>]

Hi Everyone,

I have been using Semantic Knowledge Graph for document summarization, term
correlation and document similarity. It has produced very good results
after appropriate tuning.

I was wondering if there is any way the Semantic Knowledge Graph can be
used to for Named Entity Extraction like person names, companies etc.
Related cases could be like below.

1. Extracting top named entities given a specific document or set of
documents.
2. Given a named entity (let's say person name), return top N entities
which are conceptually related to that entity.

Does anyone have an idea as to how this can be achieved? Any direction
would be a great help!

Thanks And Regards,
Pratik

Re: Named entity extraction/correlation using Semantic Knowledge Graph

[Pratik Patel <pr...@semandex.net>]

I am on look out for ideas too but I was thinking of using some NER
technique to index named entities in a specific field and then use Semantic
Knowledge Graph on that specific field i.e. limit SKG queries to that field
only. I am not sure however if this would produce desired results. I don't
have a training corpus yet. Essentially what I want is something like a
Solr Filter for entities or a request handler which can extract entities at
query time.


On Wed, Oct 17, 2018 at 4:45 PM Alexandre Rafalovitch <ar...@gmail.com>
wrote:

> Solr does have:
> 1) OpenNLP that does NER specifically
> 2) TextTagger that does gazeteer NER based on existing list but with
> Solr analysis power
>
> I would be curious to know how Semantic Knowledge Graph could be used
> from NER (or even for other things you already have used it for), but
> I am not sure it is clear what specifically you invisage. As in, is
> there training corpus, are you looking at NGram techniques, etc.
>
> Regards,
>    Alex.
> On Wed, 17 Oct 2018 at 13:40, Pratik Patel <pr...@semandex.net> wrote:
> >
> > Hi Everyone,
> >
> > I have been using Semantic Knowledge Graph for document summarization,
> term
> > correlation and document similarity. It has produced very good results
> > after appropriate tuning.
> >
> > I was wondering if there is any way the Semantic Knowledge Graph can be
> > used to for Named Entity Extraction like person names, companies etc.
> > Related cases could be like below.
> >
> > 1. Extracting top named entities given a specific document or set of
> > documents.
> > 2. Given a named entity (let's say person name), return top N entities
> > which are conceptually related to that entity.
> >
> > Does anyone have an idea as to how this can be achieved? Any direction
> > would be a great help!
> >
> > Thanks And Regards,
> > Pratik
>

Re: Named entity extraction/correlation using Semantic Knowledge Graph

[Alexandre Rafalovitch <ar...@gmail.com>]

Solr does have:
1) OpenNLP that does NER specifically
2) TextTagger that does gazeteer NER based on existing list but with
Solr analysis power

I would be curious to know how Semantic Knowledge Graph could be used
from NER (or even for other things you already have used it for), but
I am not sure it is clear what specifically you invisage. As in, is
there training corpus, are you looking at NGram techniques, etc.

Regards,
   Alex.
On Wed, 17 Oct 2018 at 13:40, Pratik Patel <pr...@semandex.net> wrote:
>
> Hi Everyone,
>
> I have been using Semantic Knowledge Graph for document summarization, term
> correlation and document similarity. It has produced very good results
> after appropriate tuning.
>
> I was wondering if there is any way the Semantic Knowledge Graph can be
> used to for Named Entity Extraction like person names, companies etc.
> Related cases could be like below.
>
> 1. Extracting top named entities given a specific document or set of
> documents.
> 2. Given a named entity (let's say person name), return top N entities
> which are conceptually related to that entity.
>
> Does anyone have an idea as to how this can be achieved? Any direction
> would be a great help!
>
> Thanks And Regards,
> Pratik