As I said: have you check the proper noun detection and POS tagging in your chain?

For instance, enhancing the text "I studied at the University of Salzburg, which is based in Austria" works at the demo server:

http://dev.iks-project.eu:8081/enhancer/chain/dbpedia-proper-noun

Here the details:

http://stanbol.apache.org/docs/trunk/components/enhancer/engines/entitylinking#proper-noun-linking-wzxhzdk14enhancerengineslinkingpropernounsstatewzxhzdk15

Cheers,


On 15/07/13 15:27, Sawhney, Tarandeep Singh wrote:
Just to add to my previous email

If i add another individual in my ontology "MyUniversity" under class
University



     <!--
http://www.semanticweb.org/vi5/ontologies/2013/6/untitled-ontology-13#MyUniversity-->

     <owl:NamedIndividual rdf:about="
http://www.semanticweb.org/vi5/ontologies/2013/6/untitled-ontology-13#MyUniversity
">
         <rdf:type rdf:resource="
http://www.semanticweb.org/vi5/ontologies/2013/6/untitled-ontology-13#University
"/>
         <rdfs:label>MyUniversity</rdfs:label>
     </owl:NamedIndividual>


So with all configurations i have mentioned in the word document (in google
drive folder), when i pass text with "MyUniversity" in it, my enhancement
chain is able to extract "MyUniversity" and link it with "University" type

But same set of configurations doesn't work with individual "University of
Salzburg"

If anyone of you please provide help on what are we missing to be able to
extract custom entities which has space in between, will be a great help to
proceed further on our journey with using and contributing to stanbol

with best regards,
tarandeep



On Mon, Jul 15, 2013 at 5:57 PM, Sawhney, Tarandeep Singh <
tsawh...@innodata.com> wrote:

Thanks Sergio and Dileepa for your responses

We haven't been able to resolve the issue. We therefore decided to keep
just one class and one instance value "University of Salzburg" in our
custom ontology and try to extract this entity and also link it but we
could not get this running. I am sure we are missing some configurations.

I am sharing a google drive folder at below link

https://drive.google.com/folderview?id=0B-vX9idwHlRtRFFOR000ZnBBOWM&usp=sharing

This folder has 3 files:

1) A word document which shows felix snapshots of what all configurations
we did while configuring Yard, yardsite, entiy linking engine and weighted
chain
2) our custom ontology
3) the result of SPARQL against our graphuri using SPARQL endpoint

May i request you all to please look at these files and let us know if we
are missing something in configurations.

We have referred to below web links in order to configure stanbol for
using our custom ontology for entity extraction and linking

http://stanbol.apache.org/docs/trunk/customvocabulary.html
http://stanbol.apache.org/docs/trunk/components/entityhub/managedsite

http://stanbol.apache.org/docs/trunk/components/enhancer/engines/entityhublinking

http://stanbol.apache.org/docs/trunk/components/enhancer/chains/weightedchain.html

Thanks in advance for your valuable help.

Best regards
tarandeep



On Sat, Jul 13, 2013 at 5:57 PM, Sergio Fernández <
sergio.fernan...@salzburgresearch.at> wrote:

Hi,

I'm not an expert on entity linking, but from my experience such
behaviour could be caused by the proper noun detection. Further details at:

http://stanbol.apache.org/**docs/trunk/components/**
enhancer/engines/entitylinking<http://stanbol.apache.org/docs/trunk/components/enhancer/engines/entitylinking>

In addition, I'd like to suggest you to take a look to the netiquette in
mailing lists. This is an open source community; therefore messages
starting with "URGENT" are not very polite. Specially sending it on Friday
afternoon, when people could be already out for weekend, or even on
vacations.

Best,
Sergio



On 12/07/13 15:54, Sethi, Keval Krishna wrote:

Hi,

I am using stanbol to extract entitiies by plugging custom vocabulary as
per 
http://stanbol.apache.org/**docs/trunk/customvocabulary.**html<http://stanbol.apache.org/docs/trunk/customvocabulary.html>

Following are the steps followed -

   Configured Clerezza Yard.
   Configured Managed Yard site.
   Updated the site by plugging ontology(containing custom entities) .
   Configured Entity hub linking Engine(*customLinkingEngine*) with
managed
site.
   Configured a customChain which uses following engine

     -  *langdetect*
     - *opennlp-sentence*
     - *opennlp-token*
     - *opennlp-pos*
     - *opennlp-chunker*
     - *customLinkingEngine*

Now, i am able to extract entities like Adidas using *customChain*.

However i am facing an issue in extracting entities which has space in
between. For example "Tommy Hilfiger".

Chain like *dbpedia-disambiguation *(which comes bundeled with stanbol
instance) is rightly extracting entities like  "Tommy Hilfiger".

I had tried configuring  *customLinkingEngine* same as *
dbpedia-disamb-linking *(configured in *dbpedia-disambiguation* ) but it
didn't work to extract above entity.

I have invested more than a week now and running out of options now

i request you to please provide help in resolving this issue


--
Sergio Fernández
Salzburg Research
+43 662 2288 318
Jakob-Haringer Strasse 5/II
A-5020 Salzburg (Austria)
http://www.salzburgresearch.at





--
Sergio Fernández
Salzburg Research
+43 662 2288 318
Jakob-Haringer Strasse 5/II
A-5020 Salzburg (Austria)
http://www.salzburgresearch.at

Reply via email to