Dear Andreas, The latest version of KIM 2.3 will give you much better performance if you haven't already upgraded. We have updated the documentation too - the online version is here - http://ontotext.com/kim/doc/sys-doc/HomePage.html and the release notes are here - http://ontotext.com/kim/release-notes.html. You can download at the usual place - http://ontotext.com/kim/KIM-downloads.html
As for your questions: 1) Often when I try to annotate documents (including standard corpus) KIM only recognizes numerical text string, dates and some Organizations mentioned in the text. Sometimes it recognizes a lot more in the text like persons etc from the same set of documents. Sometimes, it helps to clear the entities, the docs and the cache according to the “How to” section in the documentation but not always. How can this be solved? These sound like cache corruptions issues. The old releases of KIM were sensitive to abrupt shutdowns and show these exact problems if they are not stopped using the stopKIM.bat / sh script. The new version is more stable in that area. Still KIM is a server system and it's recommended to always shut it down cleanly with stopKIM 2) I have not yet seen any relations to be extracted from the text and I not sure it works at all. What kind of relations are KIM capable to recognize. Does the standard corpus contain any relations that KIM is supposed to find? I agree that the documentation is not very explicit on relation extraction capabilities. In the next few days, we will update the online documentation and I'll write you back again. In the mean time, see the what extracted relation we have on the demo server . Try this: - open http://kim.sirma.bg/KIM/screen/PredefinedEntityPatterns.jsp - Choose Predefined Entity Pattern - Person hasPosition JobPosition withinOrganization Organization - Type who has position - president - Click Documents You'll see a list of documents that contain annotations with extracted relations and you can use them as an example. The list will truncated due to performance concerns. Finally, it will be helpful if you mentioned the version of KIM that you are using. If you have a more specific issue that you need support with, send us a direct email at [EMAIL PROTECTED] with attached logs from the log folder and an example document from your set. Best regards, Marin Nozhchev KIM Platform, Lead Developer Andreas Horndal wrote: > > At a first try KIM seems to be a very capable platform for semantic > annotation. However, I have some problems. > > 1) Often when I try to annotate documents (including standard corpus) > KIM only recognizes numerical text string, dates and some > Organizations mentioned in the text. Sometimes it recognizes a lot > more in the text like persons etc from the same set of documents. > Sometimes, it helps to clear the entities, the docs and the cache > according to the “How to” section in the documentation but not always. > How can this be solved? > > 2) I have not yet seen any relations to be extracted from the text and > I not sure it works at all. What kind of relations are KIM capable to > recognize. Does the standard corpus contain any relations that KIM is > supposed to find? > > Thanks > > Andreas Horndahl > > ------------------------------------------------------------------------ > > _______________________________________________ > 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
