Dear Mircea,
thanks for the clarification.
Currently we gather the co-occurrence information in the context of the
entire document. This means that we won't be able to automatically suit
your needs at this stage.
in the near future we are planning to extend this to cover different
contexts simultaneously: sentence, paragraph, document. From these one
will be able to make any further consolidations and aggregations.
Even strange ones like Give me all the entities co-appearing in the same
sentence with New York and also are contained in documents that are
within one week window around 11th of September each year.
So discussions in this direction will be interesting for us too (if you
need someone to talk to)
Also the results of your research and development will be of interest.
have a good day
borislav
Mircea Trifan wrote:
>
> Hi Borislav,
>
> Thanks for letting me know about the CORE module. I'm trying to
> implement a KIM client application that constructs a graph of
> co-occurences of entities within text sentences and then issue user
> defined queries on top of it. I'm not trying to collect any statistics
> at this stage.
>
> Thanks+regards,
> Mircea.
>
>
>
> borislav popov wrote:
>
>> Dear Mircea,
>> it is not clear for us whether you know about the CORE module of
>> KIM or you are trying to implement co-occurrence yourself.
>> In the last installations we have included the co-occurrence and
>> ranking of entities module which is based on a relational database
>> aggregating statistics from the semantic annotation process.
>> Namely, it "counts" the occurrences of entities in documents. It can
>> be tried out on our official demo server http://ontodemo.sirma.bg/KIM
>> It can also display popularity timelines for a group of entities over
>> a set of documents.
>>
>> Can you please tell us whether you prefer to aggregate the statistics
>> yourself or you are trying to use the CORE module ?
>>
>> Regarding your question whether this method will work (explore
>> entities and gather stats) : yes this is basically the way we do it too.
>>
>> all the best
>> borislav
>>
>>
>> Mircea Trifan wrote:
>>
>>> Hi,
>>>
>>> I'm trying to build an application based on KIM that should be able
>>> to find for now co-occurrences of named entities extracted by KIM
>>> within sentences of documents in the KIM corpus. I looked in the KIM
>>> programming examples and it appears to me that I could inspect the
>>> generated annotations only at a document level granularity as per
>>> the Explore Entities section
>>> (http://www.ontotext.com/kim/doc/sys-doc/kim-api/howto-explore-entities.html#ee1).
>>>
>>> There is an example that builds a map with main label of each
>>> entity, and a count of how many times the current entity occured in
>>> the annotated document. On top of this I could build my application
>>> that could find the entities within the map in the sentences. The
>>> aliases seem to compicate a bit the problem. Does this sound right
>>> or there is a simpler way to proceed maybe using the information
>>> extraction modules?
>>>
>>> Thanks+regards,
>>> Mircea
>>>
>>> _______________________________________________
>>> NOTE: Please REPLY TO ALL to ensure that your reply reaches all
>>> members of this mailing list.
>>>
>>> KIM-discussion mailing list
>>> [email protected]
>>> http://ontotext.com/mailman/listinfo/kim-discussion_ontotext.com
>>>
>>>
>>>
>>>
>>
>
>
>
_______________________________________________
NOTE: Please REPLY TO ALL to ensure that your reply reaches all members of this
mailing list.
KIM-discussion mailing list
[email protected]
http://ontotext.com/mailman/listinfo/kim-discussion_ontotext.com