You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@stanbol.apache.org by Lukas Kahwe Smith <ml...@pooteeweet.org> on 2011/12/22 18:21:06 UTC
training stanbol
Hi,
so I have some content (legal documents) which was already manually tagged.
what i would like to do now is:
1) hand stanbol the document content and their manually assigned tags
2) hand stanbol new documents and have them tagged only using tags from the set of tags provided manually in step 1)
is that even possible?
regards,
Lukas Kahwe Smith
mls@pooteeweet.org
Re: training stanbol
Posted by Olivier Grisel <ol...@ensta.org>.
2012/2/19 Lukas Kahwe Smith <ml...@pooteeweet.org>:
>
> On Jan 11, 2012, at 09:58 , Olivier Grisel wrote:
>
>> 2011/12/22 Lukas Kahwe Smith <ml...@pooteeweet.org>:
>>> Hi,
>>>
>>> so I have some content (legal documents) which was already manually tagged.
>>>
>>> what i would like to do now is:
>>> 1) hand stanbol the document content and their manually assigned tags
>>> 2) hand stanbol new documents and have them tagged only using tags from the set of tags provided manually in step 1)
>>>
>>> is that even possible?
>>
>> I am currently working on a new module that does just that. The REST
>> API is not done yet but the backend is starting to work. You can have
>> a look at the tests here:
>>
>> https://svn.apache.org/repos/asf/incubator/stanbol/trunk/enhancer/engines/topic/src/test/java/org/apache/stanbol/enhancer/engine/topic/TopicEngineTest.java
>>
>> In particular look at the test: testTrainClassifierFromExamples
>>
>> You can follow the progress on this service / engine here:
>>
>> https://issues.apache.org/jira/browse/STANBOL-197
>
>
> How are things going?
Unfortunately I did not have the time to make big progress on
STANBOL-197 recently.
> I was hoping to revisit integration with resoltuionfinder.org in march.
Good, here is the link without the typo :) http://resolutionfinder.org/
--
Olivier
http://twitter.com/ogrisel - http://github.com/ogrisel
Re: training stanbol
Posted by Lukas Kahwe Smith <ml...@pooteeweet.org>.
On Jan 11, 2012, at 09:58 , Olivier Grisel wrote:
> 2011/12/22 Lukas Kahwe Smith <ml...@pooteeweet.org>:
>> Hi,
>>
>> so I have some content (legal documents) which was already manually tagged.
>>
>> what i would like to do now is:
>> 1) hand stanbol the document content and their manually assigned tags
>> 2) hand stanbol new documents and have them tagged only using tags from the set of tags provided manually in step 1)
>>
>> is that even possible?
>
> I am currently working on a new module that does just that. The REST
> API is not done yet but the backend is starting to work. You can have
> a look at the tests here:
>
> https://svn.apache.org/repos/asf/incubator/stanbol/trunk/enhancer/engines/topic/src/test/java/org/apache/stanbol/enhancer/engine/topic/TopicEngineTest.java
>
> In particular look at the test: testTrainClassifierFromExamples
>
> You can follow the progress on this service / engine here:
>
> https://issues.apache.org/jira/browse/STANBOL-197
How are things going?
I was hoping to revisit integration with resoltuionfinder.org in march.
regards,
Lukas Kahwe Smith
mls@pooteeweet.org
Re: training stanbol
Posted by Olivier Grisel <ol...@ensta.org>.
2011/12/22 Lukas Kahwe Smith <ml...@pooteeweet.org>:
> Hi,
>
> so I have some content (legal documents) which was already manually tagged.
>
> what i would like to do now is:
> 1) hand stanbol the document content and their manually assigned tags
> 2) hand stanbol new documents and have them tagged only using tags from the set of tags provided manually in step 1)
>
> is that even possible?
I am currently working on a new module that does just that. The REST
API is not done yet but the backend is starting to work. You can have
a look at the tests here:
https://svn.apache.org/repos/asf/incubator/stanbol/trunk/enhancer/engines/topic/src/test/java/org/apache/stanbol/enhancer/engine/topic/TopicEngineTest.java
In particular look at the test: testTrainClassifierFromExamples
You can follow the progress on this service / engine here:
https://issues.apache.org/jira/browse/STANBOL-197
--
Olivier
http://twitter.com/ogrisel - http://github.com/ogrisel
Re: training stanbol
Posted by Andreas Gruber <an...@gmail.com>.
Hi Lucas,
sorrry for the delayed answer to your question ...
You cannot do this with Stanbol exactly the way you describe it, BUT
you can achieve at least one of your goals "tagging new documents
only using tags from the set of tags provided manually".
Stanbol provides you with the ability to use custom vocabularies[1],
e.g. your list of tags organized as SKOS vocabulary[2] (or any other
RDF data).
The engine you use for generating linking suggestions will then just
point to the tags you provide.
Best,
Andreas
[1] http://incubator.apache.org/stanbol/docs/trunk/customvocabulary.html
[2] http://www.w3.org/2009/08/skos-reference/skos.html
On Thu, Dec 22, 2011 at 6:21 PM, Lukas Kahwe Smith <ml...@pooteeweet.org> wrote:
> Hi,
>
> so I have some content (legal documents) which was already manually tagged.
>
> what i would like to do now is:
> 1) hand stanbol the document content and their manually assigned tags
> 2) hand stanbol new documents and have them tagged only using tags from the set of tags provided manually in step 1)
>
> is that even possible?
>
> regards,
> Lukas Kahwe Smith
> mls@pooteeweet.org
>
>
>
--
Andreas Gruber
twitter @andresgruber
skype: andreas_gruber