Hi Alistair,
We did our checks. Let me try to explain – and if you need
further clarification – you will let us know. The resources playing a part are: - Entity-priority.conf – a configuration file defining absolute and comparative priority of/between classes/instances. An example config is attached - Com.ontotext.kim.model.EntityPriority – a class loading the config and providing API for getting priorities, thresholds and deltas of priorities - Com.ontotext.kim.gate.AnnotationFilter – a processing resource (in terms of GATE) which filters overlapping (entity) annotations. Basically it focuses on filtering out duplications and recognized entities, when there is already a “trusted” entity annotation (with instance feature filled in) at this place. It appears that in the default KIM setup – as present in our public evaluation downloads the first two (conf and entity priority class) are not used in any way. The AnnotationFilter is used instead. A hybrid of these two would prove beneficial and flexible enough in your situation – from what we know so far. To start using the entity priority class we need more than just re-configuration of the IE – we need to change the Annotation Filter or come up with a new PR. These are the options: 1. Wait a couple of months for the new major public release of KIM if you are not in a terrible pain from the situation as is. a. +: single integration; other new features and improved stability; improved installation procedure b. -: will come later 2. We plan a change request to be delivered faster if you are in a hurry a. +: faster b. –: integration effort likely to be followed by more effort after the release of the public version 3. You need it badly now: we can discuss this option as well if this is the situation – after reorganizing our team and assignments – it is possible Pick up your favorite option And have a great day! borislav From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of borislav popov Sent: Wednesday, February 13, 2008 11:28 AM To: [EMAIL PROTECTED]; [email protected] Subject: Re: [KIM-discussion] Setting precendence in types of entities extracted Hi Alistair, Nice to hear you still try out things with KIM – maybe we can work closer on this. Regarding your question – yes there is such functionality in KIM, but it is not documented in the public release. Let us have a short assessment of the state of affairs and get back to you. Regarding your second question: Yes, you can do that in several ways: 1. Remove the date/time from the configured allowed entity types in KIMConstants – not so trivial as one needs to rebuild this class. We will have a more flexible way in the next public release – i have added this into our task tracking. 2. Edit the IE application so that: a. Components recognizing dates are removed b. Annotations of a particular type are removed unconditionally in the end of the processing This involves editing the IE application and this might be a bit painful for you. An alternative way is to open main.jape and add a new phase where you would do just filtering – this will not require editing of the .gapp. Let us know how we can help you And take care borislav From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED] Sent: Tuesday, February 12, 2008 12:40 PM To: [email protected] Subject: [KIM-discussion] Setting precendence in types of entities extracted Hi, Is it possible to set a precendence in the type that is assigned to exracted entities where two or more trusted entites have the same alias. For example, say I've extended the knowledge base to include names of my projects. and I have a project called ABC. There might also be a company with the same alias and from my experience occurrances of the alias ABC in the text will be recognised as a company rather than a project. Can I tell KIM to favour entities of type project over entities of type company? I know I can use the ignore list but I was wondering if there is a way that is less explicit about certain entities. Similarly is there a way to say - don't recognise certain entities such as dates or times? Thanks, Alistair _____ From: Kings,NJ,Nick,CXR3 R Sent: 01 February 2008 11:21 To: [email protected] Cc: Duke,AK,Alistair,CXR3 R Subject: Entities not be identified in text All, We're looking to apply KIM within a domain specific search engine. So, we have added to the Proton ontology, adding an instance of an organisation ("IT Futures"), and then analysed text from a document. Other entities have been identified in that text, so Gate/Annie is working, but IT Futures has not extracted as important. What steps should we now take? Do we have to edit the jape rules? Thanks, Nick Nicholas J. Kings (Nick) Next Generation Web Research IT Futures Research Centre BT Group Chief Technology Office _____ British Telecommunications plc Registered office: 81 Newgate Street London EC1A 7AJ Registered in England no. 1800000 This electronic message contains information from British Telecommunications plc which may be privileged or confidential. The information is intended to be for the use of the individual(s) or entity named above. If you are not the intended recipient be aware that any disclosure, copying, distribution or use of the contents of this information is prohibited. If you have received this electronic message in error, please notify us by telephone or email (to the numbers or address above) immediately. Activity and use of the British Telecommunications plc email system is monitored to secure its effective operation and for other lawful business purposes. Communications using this system will also be monitored and may be recorded to secure effective operation and for other lawful business purposes. No virus found in this incoming message. Checked by AVG Free Edition. Version: 7.5.516 / Virus Database: 269.20.2/1270 - Release Date: 10/02/2008 12:21 No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.5.516 / Virus Database: 269.20.4/1275 - Release Date: 12/02/2008 15:20 No virus found in this incoming message. Checked by AVG Free Edition. Version: 7.5.516 / Virus Database: 269.20.4/1275 - Release Date: 12/02/2008 15:20 No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.5.516 / Virus Database: 269.20.4/1275 - Release Date: 12/02/2008 15:20
entity-priority.conf
Description: Binary data
_______________________________________________ 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
